Pyruvate oxidase influences the sugar utilization pattern and capsule production in Streptococcus pneumoniae.

Pyruvate oxidase is a key function in the metabolism and lifestyle of many lactic acid bacteria and its activity depends on the presence of environmental oxygen. In Streptococcus pneumoniae the protein has been suggested to play a major role in metabolism and has been implicated in virulence, oxidative stress survival and death in stationary phase. Under semi-aerobic conditions, transcriptomic and metabolite profiling analysis of a spxB mutant grown on glucose showed minor changes compared to the wild type, apart from the significant induction of two operons involved in carbohydrate uptake and processing. This induction leads to a change in the sugar utilization capabilities of the bacterium, as indicated by the analysis of the growth profiles of the D39 parent and spxB mutant on alternative carbohydrates. Metabolic analysis and growth experiments showed that inactivation of SpxB has no effect on the glucose fermentation pattern, except under aerobic conditions. More importantly, we show that mutation of spxB results in the production of increased amounts of capsule, the major virulence factor of S. pneumoniae. Part of this increase can be attributed to induction of capsule operon (cps) transcription. Therefore, we propose that S. pneumoniae utilizes pyruvate oxidase as an indirect sensor of the oxygenation of the environment, resulting in the adaption of its nutritional capability and the amount of capsule to survive in the host.

Introduction

is an important human pathogen as it is a common cause of respiratory pathologies and serious invasive diseases such as pneumonia, sepsis and meningitis; especially infants, the elderly and immuno-compromised individuals are at risk. Annually, over 1 million children die of pneumonia and meningitis and in the US alone 40.000 deaths each year are caused by pneumococcal pneumonia or meningitis [1]. Capsule, an extracellular structure of polysaccharide nature attached to the cell wall and membrane, is one of the main virulence factors of ; its main purpose is to inhibit complement-mediated opsono-phagocytosis and acapsular mutants are avirulent [2-5]. The composition of the capsule varies from strain to strain and up to 93 different capsular serotypes have been identified [6]. Capsular polysaccharide synthesis is intimately connected with central carbon metabolism as it requires activated carbohydrates; a number of these NDP-sugars derive from the glycolytic intermediates glucose 6-phosphate and fructose 6-phosphate [7-9]. A direct correlation between capsule thickness and virulence has been reported [4,10]. The amount of capsule varies throughout infection and decreases dramatically in the proximity of, and while invading, eukaryotic cells [11,12]. In addition, displays at least two phenotypic variants, transparent and opaque, that are to a large extent determined by the quantity of capsule and have differences in virulence properties [13,14]. Transparent pneumococci are more efficient colonizers of the nasopharynx, produce less capsule, but more cell-wall associated teichoic acids than opaque variants [13-15]. Additionally, the two variants show differential production of various proteins, one of which is the pneumococcal pyruvate oxidase enzyme SpxB [16]. However, the factors determining the occurrence of phase variation and the regulation of the amount of capsule are still not entirely clear, nor the connections that exist between central carbon metabolism and capsule production.

SpxB is reported to play a central role in carbohydrate metabolism under semi-aerobic (microaerobic) conditions and in particular in the generation of the phosphoryl donor metabolite acetyl-phosphate (Ac–P) [17]. The enzyme decarboxylates pyruvate into Ac–P by reducing oxygen (O2) to hydrogen peroxide (H2O2) and consuming inorganic phosphate (Pi). SpxB activity in might produce H2O2 in the mM range [17,18], concentrations higher than those generated by many other species [18,19], and sufficient to kill or inhibit other nasopharyngeal flora members, such as and [20]. Furthermore, these amounts have cytotoxic effects on human cells [21-23]. The spxB gene has also been designated as a suicide gene since the H2O2 induces pneumococcal death and reduces survival in stationary phase under semi-aerobic (microaerobic) conditions [24]. In addition, the H2O2 produced endogenously influences the membrane composition through modulation of fatty-acid biosynthesis F (FabF) protein activity [25,26]. Interestingly, and despite the potential deleterious effect of its product, SpxB has also been associated with streptococcal resistance to H2O2 [19]. Besides its function in central metabolism, a prominent role of SpxB in the virulence of the bacterium, especially in colonization and pneumonia, is corroborated by a number of studies [17,24,27,28]. Here, we studied the role of SpxB under semi-aerobic conditions as this is likely to reflect the atmospheric conditions encountered by in various host niches during infection [29]. We describe that the inactivation of spxB results in the overproduction of capsule under semi-aerobic conditions, which is partially mediated by increased transcription of the capsule locus. Furthermore, we characterized the transcriptomic and metabolic consequences of spxB inactivation, which consist of a change in sugar utilization capacities and hypothesize that these effects also contribute to its role in virulence.

Results

Inactivation of spxB leads to an increase in capsule production

While generating a transposon library in strain D39 using the pGh9:ISS1 T7 plasmid [30], a colony displaying characteristics of overproducing capsule was isolated; on solid medium this isolate yielded large, mucoidal and slimy colonies (Figure 1A). To ensure that the observed phenotype was genetically linked to the transposon insertion, the chromosomal DNA of this mutant was back-crossed into strain D39 which gave rise to the same phenotype. Transformation into strain TIGR4 also resulted in the same phenotype, indicating that the phenotype was serotype independent (Figure 1B). Comparison of the TIGR4 strain containing the insertion with an opaque and transparent variant of the same strain indicated that the phenotype of the mutant was dissimilar to the opaque phase variant as it seemed to produce even more capsule (Figure 1B). Introduction of the chromosomal DNA into an unencapsulated mutant of D39 or strain R6 [31,32], a spontaneous derivative of D39 displaying higher transformability, increased pyruvate oxidase activity and no capsule, did not result in the mucoidal phenotype (data not shown), which indicates that the observed phenotype is exclusively capsule-dependent. Excision of the pGh9:ISS1 T7 plasmid in two independent colonies of D39 resulted in the same phenotype as the original mutant, as did excision of the plasmid in a D39 strain that was backcrossed with the chromosomal DNA of the original mutant, excluding the possibility that the phenotype was due to a point mutation in the original mutant. Determination of the insertion site showed that the ISS1 element was inserted into the 3’ end of the spxB gene.

Figure 1

Inactivation of spxB led to an increase in capsule production.

(A) Phenotype on Glc-M17 agar plate of D39 (D39), D39 containing the pGh9 T7 plasmid (D39pGh9 T7) and D39 in which the plasmid is excised leaving the ISS1 element in spxB (D39spxB). (B) Phenotype on Glc-M17 agar plate of TIGR4 transparent variant (T4 transparent), TIGR4 opaque variant (T4 opaque) and TIGR4 in which the pGh9 T7 plasmid is inserted into the genome (T4 pGh9 T7). (C–F) TEM pictures of grown in broth to exponential phase and stained with LRR. (C) D39, 9700 times magnified; (D) D39spxB 9700 times magnified; (E) D39, 135.000 times magnified; (F) D39spxB 135.000 times magnified.

The spxB gene encodes the pyruvate oxidase, which generates Ac–P and H2O2 and whose activity plays a key role in pneumococcal colonization and pathogenesis [17,24,27,28]. The amount of H2O2 produced by the mutant was reduced at least 80% compared to the wild type and PCR analysis showed the insertion of approximately 600-bp extra into the spxB gene in all excised mutants (data not shown) and no SpxB protein could be detected (Figure S1), confirming that the insertion of ISS1 disrupted spxB and its function.

The observed phenotype of the colonies (large, mucoidal and slimy) could be due to the reported absence of death in stationary phase of spxB mutants and the ensuing increase in biomass production over time [24] or due to increased capsule production. Therefore, the amount of capsule was determined in both the wild type and the spxB mutant (D39spxB) in late-exponential (Figure 2A) and early-stationary phases of growth (Figure 2B) in Glc-CDM. The data showed that the amount of capsule in the D39spxB mutant significantly increased by circa 30% relative to the wild type levels (students t-test P≤0.01), meaning that the observed larger colonies are to some extent due to increased capsule amounts rather than differences in survival. Furthermore, TEM pictures of bacteria grown to exponential phase in liquid medium, also indicated an increased amount of capsule in the mutant (Figure 1C, 1D) as well as some subtle changes in the cell membrane and or peptidoglycan (Figure 1E, 1F).

Figure 2

Effect of spxB mutation on capsule production.

Estimation of capsule was performed based on the determination of its glucuronic acid content in strains D39 (white bars), D39spxB (dark grey bars) and D39spxB (complemented strain) (light grey bars) in late-exponential (A) and early-stationary (B) phases of growth. Cultures were grown in CDM containing 1% (wt/vol) Glc at 37°C, with pH control (6.5), and under semi-aerobic conditions (for details see Materials & Methods). Determinations were done twice in three independent cultures and the values are means ± SD. * = P≤0.01 students t-test. For D39spxB , the capsule was determined twice for each of two independent cultures.

In order to determine that the observed phenotype was indeed due to the inactivation of spxB, complementation of the spxB mutation was performed, which restored the production of SpxB to levels comparable to that of the wild type (Figure S1). The levels of H2O2 measured (41 ± 4 µM and 80 ± 22 µM at late-exponential and early-stationary phases of growth, respectively) in the complemented strain, indicate restoration of SpxB activity. Importantly, capsule amounts were restored back to the wild type levels in the complemented strain (D39spxB ) (Figure 2). Thus, in the conditions studied our results show that inactivation of spxB increased the amount of capsule in . In this light, a role of the gene downstream of spxB, spd0637, in the appearance of the phenotype can be ruled out. Mutations of spxB or reduction in SpxB activity resulting in mucoid and or opaque-like appearances is a recurrent observation [18,28,31,32].

Transcription of capsule (cps) operon is increased in the spxB mutant

Next we determined whether the increase in capsule production was mediated by changes in transcription of the cps operon by determining the activity of the cps promoter, localized upstream of the first gene, cpsA. A fusion with a promoterless lacZ was generated in pORI13 [33] and introduced into D39 and the D39spxB mutant. Both in mid-exponential phase as well as late-exponential phase, transcription of the cps promoter was increased in the D39spxB mutant by about 45% relative to the wild type cps transcriptional levels in Glc-CDM, which was statistically significant in each case (Figure 3A). Thus, the increased capsule production is at least partially mediated by a change in transcription of the capsule locus. To investigate whether the increased transcription of cps was due to oxidative stress induced by external H2O2 or the metabolic activity of SpxB, D39 and the D39spxB mutant were grown with and without 200 U/ml bovine liver catalase in BHI. Cps transcription in the wild type was not increased to the level of the spxB mutant by the detoxification of H2O2 indicating that the effect is not mediated by the presence of this reactive oxygen species (Figure 3B). Determination of cps promoter transcriptional levels of D39 and spxB in the presence of catalase was peformed by growing the cells in a complex medium (BHI) (Figure 3B) rather than CDM (Figure 3A). The difference in medium composition might explain the lower induction ratio (about 15%) of cps transcription in the spxB mutant observed in BHI (Figure 3B) as compared to that in CDM (45% induction) (Figure 3A).

Figure 3

Effect of spxB mutation and addition of catalase on cps transcription.

Transcription of the cps promoter was estimated by measuring the β-galactosidase activities of strains D39 (bars with white background), D39spxB (bars with grey background) and D39spxB and its parent in the presence of 200 U/ml bovine liver catalase (spotted bars) harbouring pORI Pcps (see Table 1) in mid-exponential (A, left panel and B) and late-exponential (A, right panel) phases of growth. Cultures were grown in CDM containing 1% (wt/vol) Glc (A) or in BHI (B) at 37°C, and under semi-aerobic conditions. All the determinations were done at least in triplicate and the values are means ± SD. a.u. = arbitrary units.

Table 1

Genes differentially regulated in the D39spxB mutant grown in Glc-M17.

SPnumber	SPDnumber	TIGRannotation	Foldchange	Bayes.p
SP0095	SPD0091	conserved hypothetical protein	0.4	8.38E-08
SP0317	SPD0289	4-hydroxy-2-oxoglutarate aldolase/2-deydro-3-deoxyphosphogluconate aldolase	7.4	3.00E-15
SP0318	SPD0290	carbohydrate kinase, PfkB family	9.5	0
SP0319	SPD0291	conserved domain protein	9.6	0
SP0320	SPD0292	oxidoreductase, short chain dehydrogenase/reductase	8.6	0
SP0321	SPD0293	PTS system, IIA component	2.3	7.08E-07
SP0322	SPD0294	glucuronyl hydrolase	3.5	4.66E-12
SP0323	SPD0295	PTS system, IIB component	2.6	4.30E-07
SP0324	SPD0296	PTS system, IIC component	4.6	4.54E-09
SP0325	SPD0297	PTS system, IID component	10.2	0
SP0506		integrase/recombinase, phage integrase family	2.4	1.67E-06
SP0626	SPD0546	branched-chain amino acid transport system II carrier	0.3	3.61E-11
SP0731	SPD0637	conserved domain protein	0.3	2.22E-16

Genes were considered differentially expressed when the fold change was ≥ 2, ≤ 0.5 with a Bayes p ≤ 0.00001

Deletion of spxB has only a minor influence on the transcriptome of D39

To obtain a better insight into why inactivation of spxB resulted in an increase in capsule production, the transcriptome of the spxB mutant was compared to that of the wild type strain under semi-aerobic conditions (Table 1). Interestingly, inactivation of spxB had only a modest influence on the transcriptome of (Table 1). However, in line with its role in metabolism [19,32,34], two divergent putative operons that are likely involved in carbohydrate metabolism were highly induced (Table 1). The first one (spd0289 to spd0292) contains genes whose homology indicates involvement in the catabolism of ketogluconate [35]. The second putative operon encodes four genes (spd0293, spd0295-7) showing homology to a mannose-family PTS system that is predicted, based on homology of the IIC and IID components, to use as substrate N-acetylglucosamine (TransportDB, ), and one gene, spd0294 encoding a glucuronyl hydrolase [36]. A recent report provided evidence for the involvement of the Man-family PTS in the utilization of hyaluronic acid, a glycosaminoglycan consisting of repeating disaccharides, which are composed of D-glucuronic acid and D–N-acetylglucosamine [37]. Contrary to the expectations, no changes were observed in genes of the capsule locus. This could be due to the modest increase in capsule amount, below 2 fold, which was the cut off value used.

It should be noted that the insertion of the ISS1 element at the 3’ end of the spxB gene yielded no changes in the transcription level of the gene. Although spxB and its downstream gene spd0637 do not appear to be located in an operon, expression of spd0637 was significantly reduced in the mutant. Theoretically, this could also be the cause of the observed effects of the ISS1 insertion into spxB. However, complementation of the spxB mutation reverted the phenotype of the colonies back to the wild type appearance, indicating that spd0637 repression played no role in the increase in capsule production.

Growth and fermentation profiles are not affected by spxB inactivation under semi-aerobic conditions

Given the metabolic function of SpxB and in order to explore a potential connection between carbohydrate metabolism and capsule production, growth properties of D39spxB and its parent strain D39 were determined in Glc-M17 as this was the medium in which the original phenotype was observed. Under these conditions, the spxB mutation had no significant effect on the growth profile or fermentation type of D39 (Figure 4A and Table 2). Both D39 and its spxB derivative grew to maximal OD600 of about 1.5 in 6 h at similar specific growth rates (Table 2), showing a typical homolactic fermentation pattern (Table 2), with lactate accounting for about 90% of the glucose consumed. Considering the absence of major differences in growth characteristics of strain D39 and its D39spxB mutant in the complex medium M17, we wondered whether in a chemically defined medium (CDM) the loss of the metabolic enzyme SpxB would have a more pronounced effect. Furthermore, the use of CDM not only facilitates the interpretation of results, but also allows the use of NMR techniques for metabolic characterization.

Figure 4

Growth profiles of strains D39, D39spxB and D39spxB under semi-aerobic conditions.

Cultures were grown in A) M17 without pH control (initial pH 6.5) or B) CDM with pH control (pH 6.5), with 1% (wt/vol) Glc at 37°C, in both cases under semi-aerobic conditions (for details see Materials and Methods). The plotted growth curves are from a representative experiment. For each condition at least three independent experiments were performed, except for the complemented strain (D39spxB ) which was performed twice, and the error was below 15%. Symbols: (triangles), D39spxB; (squares), D39; (diamonds), D39spxB . Arrows in the CDM grown cultures indicate sampling time for capsule determination and intracellular metabolite analysis (Figures 2 and 5).

Table 2

Effect of spxB deletion on the end-product yields and growth and energetic parameters of D39 grown in Glc-M17 and Glc-CDM under semi-aerobic conditions[a].

	Semi-aerobiosis
	Glc-M17		Glc-CDM
	D39spxB	D39	D39spxB	D39
Product yields[b]
Lactate	1.75 ± 0.06	1.78 ± 0.02	1.69 ± 0.03	1.63 ± 0.01
Formate	0.05 ± 0.05	0.10 ± 0.03	0.08 ± 0.01	0.06 ± 0.01
Pyruvate[c]			0.09 ± 0.02	0.04 ± 0.01
Acetate	BDL	BDL	BDL	BDL
H2O2 (µM)	ND	ND	BDL	10 ± 0
Consumed substrate (%)	36 ± 2	35 ± 1	28 ± 1	28 ± 0
Carbon balance[d]	88 ± 1	89 ± 1	93 ± 1	87 ± 1
Redox balance[e]	84 ± 3	83 ± 1	88 ± 1	85 ± 1
Biomass yield (g mol-1 of Glc)	24.8 ± 0.4	23.7 ± 0.2	24.9 ± 3.1	26.2 ± 1.0
ATP yield (mol mol-1 of Glc)	1.8 ± 0.0	1.8 ± 0.1	1.8 ± 0.1	1.7± 0.0
YATP (g of biomass mol-1 of ATP)	14.1 ± 0.2	13.3 ± 0.1	14.2 ± 1.3	15.0 ± 0.7
µmax (h-1)	0.85 ± 0.01	0.87 ± 0.06	0.70 ± 0.05	0.83 ± 0.04

D39 and D39spxB were grown in complex medium (M17) or chemically defined medium (CDM) containing 1% (wt/vol) glucose, under semi-aerobic conditions as described in materials and methods. Product yields and growth and energetic parameters were determined from substrate and fermentation product analysis by HPLC at the time-point of maximal biomass achieved by both strains;

Product yields: [End-product, mM]/[Glucose consumed, mM];

Blank cells, negative yields were found for these conditions (cells used pyruvate from the medium);

Carbon balance is the percentage of carbon in metabolized Glc that is recovered in the fermentation products (lactate and formate) and pyruvate;

Redox balance is the ratio between [lactate] and 2 x [glucose] multiplied by 100;

Dry weight (DW) was used as a measure of cell mass. ND, not determined. BDL, below detection limit;

Values of at least two independent experiments were averaged and errors are reported as ± SD.

Strains D39 and D39spxB were grown in Glc-CDM (Figure 4B, Table 2). The specific growth rates and maximal OD600 were lower in Glc-CDM than in the complex medium (Glc-M17). In the defined medium D39 and D39spxB displayed identical growth properties, with respect to the biomass yield, lactate and formate yields, and bioenergetic parameters (Table 2). Strain D39 presented, however, a maximal growth rate significantly higher (P value of 0.002) than D39spxB. Consistent with inactivation of pyruvate oxidase, the pyruvate yield was 2-fold higher in the D39spxB mutant (Table 2). Also in agreement, H2O2 (about 10 µM) was detected in the extracellular medium during the growth of the parent strain, but not in D39spxB. These data corroborate the loss of SpxB function in our mutant strain. It is plausible to assume that the low H2O2 concentration measured under semi-aerobic conditions arose from a combination of factors: the low dissolved oxygen and H2O2 decomposition by the action of protection mechanisms against its toxicity. Indeed, we are not the first to report lower H2O2 production in than that expected from stoichiometry [34]. Curiously, Glc was still abundant when growth ceased (70% of the initial Glc remained in the culture medium, Table 2). Strains D39 and D39spxB were grown under controlled neutral pH, thus growth arrest was likely due to a nutritional limitation other than Glc. Overall, spxB mutation slightly decreases the growth rate of D39 under semi-aerobic conditions and increases pyruvate accumulation during Glc metabolism, but does not cause a major change in the fermentation profile.

Inactivation of spxB affects the pools of intracellular metabolites

Our data showed that the spxB mutation had a modest impact on growth and fermentation products in cultures under semi-aerobic conditions, but clearly affected the production of the polysaccharide capsule (Figure 2). The disparate capsule production could suggest an altered intracellular metabolism. In line, the specific growth rate of D39spxB was slightly, but consistently, lower than that of D39. To verify this hypothesis we measured the intracellular metabolites at the partitioning node between glycolysis and the polysaccharide 2 biosynthetic pathway, as well as several NDP-sugars, including the capsule precursors UDP-Glc and UDP-GlcUA (Figure 5). Metabolite concentrations were determined in ethanol extracts obtained during late-exponential and early-stationary phases of growth by 31P-NMR (see the time-points indicated by the arrows in Figure 4B). Interestingly, the loss of spxB led to a considerable reduction in the levels of the upper glycolytic metabolites glucose 6-phosphate (G6P) and fructose 1,6-bisphosphate (FBP) in late-exponential phase (Figure 5A). G6P is a key metabolite at the hub of glycolysis and several biosynthetic pathways, and its conversion to α-glucose 1-phosphate (α-G1P) is the first step that commits it to the synthesis of many structural polysaccharides, including type 2 polysaccharide (Figure 6). The intracellular concentration of α-G1P was similar in D39 and its D39spxB mutant. The same was true for the UDP-activated capsule precursors, UDP-Glc and UDP-GlcUA (Figure 5). These data indicate a redirection of carbon away from glycolysis in the D39spxB, and is consistent with the higher capsule production. In addition, no major differences were detected between the levels of other NDP-activated metabolites in D39 and the D39spxB mutant. As expected, the product of SpxB activity, the high energy phosphoryl donor acetyl-P (Ac-P), was reduced in the mutant by 65% of the parent strain (Figure 5), which is in the same range as the values recently reported by Ramos-Montañez et al. [32]. Similarly, in the late-exponential phase of growth, the ATP levels were not affected by the loss of SpxB activity (2.3 ± 0.3 and 2.1 ± 0.2 nmol mg DW-1, respectively). In general, the intracellular concentrations of phosphorylated metabolites in both strains were lower in early-stationary phase, except for phosphoenolpyruvate (PEP), which increased slightly.

Figure 5

Effect of spxB mutation on intracellular phosphorylated metabolites during growth in Glc-CDM under semi-aerobic conditions.

Phosphorylated metabolites were measured by 31P-NMR in ethanol extracts of D39 and D39spxB cells grown to late-exponential (LExp) (A, B) and early-stationary (EStat) (C, D) phases of growth in CDM supplemented with 1% (wt/vol) Glc as in Figure 4B. The results are averages of three to four independent growths and the average accuracy was ± 15%. Symbols: (dark grey bars), D39, (light grey bars), D39spxB. FBP, fructose 1,6-bisphosphate; G6P, glucose 6-phosphate; PEP, phosphoenolpyruvate; α-G1P, α-glucose 1-phosphate; Ac–P, acetyl-phosphate; CDP-Cho, CDP-choline; UDP-GlcUA, UDP-glucuronate; UDP-Glc, UDP-glucose; UDP-GlcNAc, UDP-N-acetylglucosamine; UDP-MurNAcp-Pep, UDP-N-acetylmuramoyl-pentapeptide; UDP-GlcN, UDP-glucosamine. LExp, late-exponential phase of growth; EStat, early-stationary phase of growth; A and C, phosphomonoesters; B and D, diphosphodiesters; (*), quantification was impaired by overlapping resonances in NMR spectra.

Figure 6

Model of the spxB deletion effect on glucose metabolism and capsule production in .

Glucose is converted by the conventional glycolytic pathway to pyruvate. Fermentation products result from the action of different competing enzymes (lactate dehydrogenase, pyruvate formate-lyase, pyruvate oxidase and hypothetically the pyruvate dehydrogenase complex). Overall, our data show that SpxB represses carbohydrate specific pathways (ketogluconate and hyaluronic acid utilization), capsule production, and stimulates formation of acetyl-P and acetate at the expense of pyruvate and lactate. Pyruvate and end-products are represented by three-dimensional bars and the height of the bar represents the relative amount; grey slashed arrows, function not experimentally verified; boxed metabolites indicate higher metabolite accumulation. Thick pointed-line, repression; Thick black arrow, activation. KDG, ketogluconate; HA, hyaluronic acid; GlcNAc, N-acetylglucosamine, GlcUA, glucuronic acid; Ac–P, acetyl-phosphate; α-G1P, α-glucose 1-phosphate; G6P, glucose 6-phosphate; FBP, fructose 1,6-bisphosphate; PEP, phosphoenolpyruvate.

SpxB skews the sugar utilization profile of

The fact that the majority of transcriptomic differences between D39 and D39spxB were observed for genes potentially involved in central metabolism, and in particular sugar catabolic pathways, prompted us to analyze the growth of the two strains on selected sugars. Genome annotation and homology studies suggested that the genes differentially expressed might play roles in the catabolism of N-acetylglucosamine (GlcNAc), glucuronic acid (GlcUA) and/or gluconic acid (GlcU). Since loss of spxB leads to higher amount of capsule polysaccharide, and considering the higher expression of a putative glucuronyl hydrolase (spd0294) and the PTS system needed to import its products [37] we surmised an aptitude of the D39spxB strain to utilize type 2 polysaccharide derived sugars (GlcUA, Glc, and rhamnose, Rha) as carbon sources. Thus, rhamnose and a sugar mixture mimicking the type 2 capsule (1 GlcUA: 2 Glc: 3 Rha) were also tested. Growth was performed in CDM supplemented with 0.25% (wt/vol) sugar in 96-well microtiter plates (Figure 7), under semi-aerobic conditions. As expected, on Glc D39 showed a modest, but significant (P value < 0.05) higher specific growth rate than the D39spxB strain (1.18 ± 0.06 as compared to 0.91 ± 0.06 h-1). The maximal OD595 was similar for both strains on Glc (about 1.1). Growth on GlcNAc was characterized by identical specific growth rates (around 0.83 h-1) for both strains, but lack of spxB resulted in 25% increase in the final biomass (from OD595 of 0.75 to 1.03). This result supports a role of the PTS system (spd0293-0297), which is differentially expressed in our transcriptomic analysis, in the catabolism of GlcNAc. In the sugar mixture, both strains displayed identical specific growth rates and maximal biomass; however, while the OD595 values of D39 decreased upon growth arrest, the D39spxB mutant showed a slow but steady increase of the OD595 values during the time of monitoring (24 h). Also, the D39spxB strain showed residual growth on GlcUA, Rha and GlcU, while D39 was totally unable to use these sugars as sole carbon sources for growth. These data strongly indicate that spxB considerably influences the nutritional utilization capabilities of .

Figure 7

Effect of spxB mutation on the ability to grow on specified sugars.

Growth profiles of D39 (A) and its derivative D39spxB (B) in CDM containing 0.25% (wt/vol) of the specified sugar. Cultures were prepared in 250 µl in 96-well microtiter plates and growth monitored at 595 nm and 37°C. Symbols: (black diamonds), Glc; (grey diamonds), GlcUA; (grey triangles), Rha; (grey circles), GlcA; (black squares), GlcNAc; (white triangles) 1 GlcUA : 2 Glc: 3 Rha. Zoomed areas: expansions of time-points 5 to 20 h for growths in rhamnose, gluconic acid and glucuronic acid. Each point of the growth curves is an average of triplicate experiments each consisting of two independent cultures, and the error was in all cases below 20%, except for rhamnose which was below 30%.

Oxygen affects differently end-product distribution in strain D39 and spxB mutant under aerobic conditions

Most studies on SpxB were performed under semi-aerobic conditions and showed a modest effect of SpxB on metabolism. However, molecular oxygen is the substrate for the SpxB catalyzed reaction. Hence, to further explore the role of SpxB activity in pneumococcal carbohydrate metabolism under highly oxygenated conditions that may occur during transmission or on the surface of the mucosa, we grew strain D39 and its derivative spxB mutant under a continuous specific air tension of 40% (constant dissolved oxygen concentration of 90 µM at 37°C). We expected visible changes in end-product profiles in strain D39, but minor or no changes in D39spxB. Indeed, the increase in oxygen availability was translated into a substantial rise in H2O2 production to about 2 mM in the parent strain D39. The concentration of hydrogen peroxide in supernatants of strain D39spxB was in the micromolar range (2 µM), three order of magnitude lower than in D39 (Figure 8 and Table 3). Under aerobic conditions, the maximal biomass reached for strains D39 and D39spxB was 0.49 ± 0.03 (time-point 3h) and 1.64 ± 0.13 (time-point 8h), respectively (Figure 8). At these time-points the D39spxB mutant consumed approximately 50% of the Glc supplied and the control strain only about 4% (Table 3). Under these conditions, the low biomass generated by D39 is most likely a consequence of high H2O2 accumulation due to SpxB activity. Loss of spxB caused a significant decrease (about 20%) in the maximal growth rate (Table 3), which suggests that SpxB activity stimulates faster growth on Glc under highly oxygenated conditions.

Figure 8

Growth and fermentation profiles in D39spxB and its parent strain D39 under aerobic conditions.

Growth curves, substrate consumption and end-products formed by D39 (A) and D39spxB (B) strains growing on Glc-CDM, at 37°C, with pH control (6.5), and under aerobic conditions (for details see supplemental Materials and Methods S1). Culture supernatant samples for substrate and end-product analysis by HPLC and/or 1H-NMR were harvested at denominated time-points along growth (bars in the plots). The plotted growth curves, substrate consumption curves and end-products bars are from a representative experiment. For each condition at least two independent experiments were performed. For the growth and substrate consumption curves the error in each point was below 10%. For the end-products concentrations, the error was below 7% for major products (> 3.5 mM) and 30% for minor products (< 3.5 mM) Symbols: (white circles), substrate consumption; (black squares), growth curve; white bars, lactate; dark grey bars, pyruvate; light grey bars, acetate; stripped bars, hydrogen peroxide. The data on D39 is also presented in [67].

Table 3

Effect of spxB deletion on the end-product yields and growth and energetic parameters of D39 grown in Glc-CDM under aerobic conditions[a].

	Aerobiosis
	Glc-CDM
	D39spxB	D39
Product yields[b]
Lactate	1.47 ± 0.02	0.47 ± 0.09
Formate	BDL	BDL
Pyruvate[c]	0.16 ± 0.01
Acetate	0.11 ± 0.00	1.53 ± 0.09
H2O2	BDL	0.82 ± 0.01
Consumed substrate (%)	49 ± 0	4.4 ± 0.6
Carbon balance[d]	87 ± 2	100 ± 0
Redox balance	ND	ND
Biomass yield (g mol-1 of Glc)	20.9 ± 2.6	68.1 ± 2.4
ATP yield (mol mol-1 of Glc)	1.9 ± 0.0	3.5 ± 0.1
YATP (g of biomass mol-1 of ATP)	11.3 ± 1.1	19.3 ± 1.1
µmax (h-1)	0.82 ± 0.03	1.00 ± 0.05

D39 and D39spxB were grown in chemically defined medium (CDM) containing 1% (wt/vol) glucose, under aerobic conditions (constant air tension of 40%) as described in materials and methods. Product yields and growth and energetic parameters were determined from substrate and fermentation product analysis by HPLC at the time-point of maximal biomass achieved by both strains;

Product yields: [End-product, mM]/[Glucose consumed, mM];

Blank cells, negative yields were found for these conditions (cells used pyruvate from the medium);

Carbon balance is the percentage of carbon in metabolized Glc that is recovered in the fermentation products (lactate and acetate) and pyruvate; Dry weight (DW) was used as a measure of cell mass. ND, not determined. BDL, below detection limit;

Values of at least two independent experiments were averaged and errors are reported as ± SD.

Before reaching stationary phase D39 produced in continuous increments the end-products acetate, H2O2 and lactate, from Glc, which reached at the time-point of growth arrest (maximal biomass) concentrations of 4.2 ± 0.1 mM, 2.2 ± 0.1 mM and 1.3 ± 0.3 mM, respectively. Interestingly, in the stationary phase the levels of acetate and H2O2 increased at the expense of lactate, suggesting the occurrence of lactate oxidase activity in this condition (Figure 8). As expected, loss of spxB led to a typical homolactic behavior, with lactate as major end-product (maximal concentration 44 ± 0 mM), accounting for 75% of the Glc consumed (Table 3); acetate was also formed in minor amounts and the yield of pyruvate was 2-fold enhanced as compared to semi-aerobic conditions. No formate was observed, suggesting suppression of pyruvate formate-lyase (PFL) activity when the oxygen concentration is kept steady around 0.09 mM. In the D39spxB mutant, the H2O2 detected was below 10 µM. Detection of acetate in the spxB mutant while spxB is inactivated and PFL suppressed by O2 indicates an alternative pathway for the formation of this organic acid. In summary, our data show that the metabolic response to highly oxygenated conditions is largely mediated by SpxB, since growth and fermentation profiles of strain D39spxB were barely affected under aerobic conditions.

Discussion

The role of pyruvate oxidase (SpxB) in the aerobic metabolism of several Gram-positive bacteria is well established [38-41]. In the human pathogen , a number of studies describe the involvement of SpxB in virulence [17,20,27], fermentation [34] and colony phenotypic variation [16,18,31,42]. However, in depth investigation of the molecular mechanisms underlying these modes of action is far from complete.

In this study we show that inactivation of spxB leads to an increase in capsule production. High capsule amounts have been consistently associated with opaque phase variants, which are interestingly also characterized by low SpxB activity [13,14,16]. However, a comparison of phenotypes on plate showed that, although displaying similar properties, the spxB mutant is not an opaque variant. Nevertheless, the observation that deletion of spxB or a reduction in SpxB activity leads to an opaque-like or mucoid phenotype on plate is recurrent [16,28,31,32]. Additionally, other disruptions in pathways leading to acetate and ATP production also seem to interfere with capsule biosynthesis, as suppressor mutations of ackA mutations often resulted in an inactivation of both spxB and capsule production [32]. Our data show for the first time, that the mucoid phenotype appearance of D39 colonies lacking SpxB activity is due to an increase in capsule biosynthesis, as indicated by the higher expression levels of the cps promoter and glucuronic acid amounts, and is independent of differences in growth or death rate.

The presence of capsule in is required for both colonization and invasive disease [3,4,10,15,45]. Too much capsule is detrimental for colonization. On the other hand, a thick capsule is optimal for invasive disease, as demonstrated by comparing transparent and opaque variants [15]. Interestingly, a number of studies report that deletion of spxB promotes invasive disease in bacteremia models [27] and leads to a defect in nasopharyngeal colonization [17,24,27,28]. The data from our study suggests that this behaviour can partially be due to the increased amounts of capsule present in the D39spxB. In concordance with this idea, spxB expression is higher in the nasopharynx than in the lungs and the blood [43], probably ensuring that not too much capsule is being produced. Also, spxB is higher expressed in transparent variants than in opaque ones, which are poor colonizers [16,44]. It should be noted that although our data show modest changes in capsule amounts between D39 and D39spxB, we believe that they are relevant in vivo as even relatively small differences in the amounts of capsule can be critical for the expression of virulence [13]. Thus, we would like to propose that SpxB activity in response to the oxygen concentration of its environment is an important signal for to determine the amount of capsule that needs to be produced. This hypothesis is strengthened by the observation that changes in oxygen seem to induces changes in capsule production in opaque variants [29]. Furthermore, transparent variants were predominantly isolated from the nasopharynx and opaque variants from blood [29]. Growth of D39 under anaerobic conditions also led to a uniform mucoid appearance of the colonies on blood agar plates and capsule amounts similar to those of D39spxB in chemically defined medium (data not shown). Additionally, growth of the bacteria in the presence of catalase did not increase cps transcription in D39, indicating that the effect is specific for SpxB activity and not by oxidative stress brought about by external H2O2. How this signaling would occur is currently unclear, one possibility is that the changes in the membrane induced by SpxB activity [25,26] mediate alterations in the production and expression of the capsule locus. Capsule synthesis is tightly interconnected with peptidoglycan synthesis and is partly regulated by membrane proteins [45], thus it is likely that disturbances in the membrane or cell wall influence capsule synthesis and vice versa. Although not conclusive, the high magnification TEM pictures of the D39spxB mutant suggest subtle changes in the membrane and or peptidoglycan, which could be mediated by spxB inactivation and thus result in a change in capsule production.

In other streptococci, CcpA, the main regulator of carbon metabolism in [46] and other Gram-positive bacteria, represses spxB [47,48]. Although there is no cre site in the promoter and the gene was also not found in extensive transcriptome analysis [46], we did attempt to generate a ΔccpA mutant in the D39spxB background. This failed unless the ccpA gene was already ectopically present in the D39spxB mutant, indicating that this combination is lethal for (data not shown). Thus, in CcpA does not regulate spxB, but repression (or activation) of a product of the CcpA regulon does seem necessary to counteract deleterious effects of spxB inactivation.

Another alternative to mediate this regulation is SpxR, however no capsule genes were differentially regulated in the transcriptome analysis of this mutant, making it a less likely candidate [28]. It is worth mentioning that while Ramos-Montañez and co-workers used D39 strain NCTC7466 as genetic background, in our work we have characterized a strain with characteristics of a Lilly derivative, at least in what concerns pyruvate oxidase activity [49]. Lilly derivatives show a considerably lower SpxB activity than NCTC strains [28]. Thus, a direct extrapolation from previous work performed with D39 NCTC might not be possible. Interestingly, transcriptome analysis of our D39spxB mutant showed more differentially transcribed genes than those reported for a spxB mutant in the NCTC isolate [28]. This could be due to allelic variance between NCTC and our isolate, differences in sampling time (phase of growth) and or in the composition of the medium used for growth. The transcriptome analysis of D39spxB showed minor changes (13 altered transcripts) compared to the wild type, despite the fact that the experiments were carried out under semi-aerobic conditions and harvest took place during early-exponential growth phase when oxygen is still thought to be present and SpxB activity can be detected. Loss of spxB did not affect the transcript levels of regulatory proteins. These findings reinforce the idea that neither SpxR nor CcpA play prominent roles in the signaling pathways mediating the SpxB phenotypes under these conditions.

Of the genes that were differentially expressed, 77% encode functions involved in central metabolism (Table 1). Of these, 9 genes organized in two transcriptional units (spd0289-92, spd0293-97) were highly induced in the spxB mutant. spd0293-97 encodes a Man-family PTS system and a glucuronyl hydrolase, which have recently been implicated in the metabolism of hyaluronic acid, and in particular in the uptake and degradation of GlcNAc-GlcUA disaccharides [37,50]. The four genes in spd0289-92 are most likely involved in the degradation of ketogluconic acid, as judged by homology comparisons [35]. Thus, a role for SpxB was surmised on the utilization of carbon sources for growth, which was corroborated by the better growth on GlcNAc compared to the wild type and the observed residual growth of the D39spxB strain on sugars potentially present in capsules or host glycans (Figure 7). Growing evidence supports the utilization of host glycans as nutritional sources for pneumococcal growth [37,51,52]. To our knowledge, we are the first to show a link between SpxB and carbohydrate utilization. Based on our results, it is tempting to hypothesize that in specific host microenvironments with oxygen where glucose becomes the major carbon source available, for instance during inflammation or hyperglycemia (e.g. lungs during infection) [53], SpxB activity minimizes the use of alternative sugar utilization pathways (repression of specific catabolic steps involved in the metabolism of sugars other than glucose) in a cost effective manner, which will allow for rapid proliferation and fast energy production, i.e. improved fitness (Figure 6).

We found that under semi-aerobic conditions, SpxB slightly stimulated the pneumococcal rate of growth on glucose and inhibited capsule production (Figure 6 and Table 2, Figures 2 and 3). Importantly, the soluble activated precursors for serotype 2 capsule synthesis are all derivatives of α-glucose 1-phosphate, which originates from the action of α-phosphoglucomutase on the glycolytic intermediate glucose 6-phosphate. The lower accumulation of glucose 6-phosphate and fructose 1,6-bisphosphate in the D39spxB mutant indicate a redirection of carbon towards biosynthetic pathways, in particular capsule (Figure 6). The data on intracellular metabolites is therefore in full agreement with our findings in respect to capsule production and growth rate, and allows us to propose that the spxB mutation affects the intracellular fluxes of .

Curiously, under the imposed semi-aerobic conditions the fermentation pattern of the pneumococcus was only marginally affected by loss of SpxB, the major difference being a higher accumulation of pyruvate in the mutant (Figure 6). This difference is consistent with a decreased level of Ac–P in D39spxB. In bacteria, a role of the central metabolite Ac–P in linking the nutritional status to global signaling, namely by acting as a phosphoryl donor to two-component response regulators has been proposed [54]. One could argue that similar molecular mechanisms could be in effect in , i.e., Ac–P phosphorylation of response regulators. However, in a recent study, Ramos-Montañez and co-workers showed that of three response regulators only WalRK had changes in expression correlating with the amounts of Ac–P; most of their work did not support phosphorylation of regulators by Ac–P as a general signaling mechanism in [32]. This is in line with the minimal effect of spxB deletion on the transcriptome, inferred in our and other studies. Rather than relaying the metabolic state to regulatory pathways, Ac–P is a major energy reservoir for ATP synthesis. In Ramos-Montañez et al. [32], the amount of ATP was not affected by spxB deletion. Likewise, in our study loss of spxB did not affect the steady-state ATP amount (2.3 ± 0.3 and 2.1 ± 0.2 nmol mg DW-1 in exponential cells of D39 and D39spxB, respectively).

It has been suggested that maintenance of wild type ATP levels is essential to cope with H2O2 imposed stress [19,32]. Our observation that the level of H2O2 produced under semi-aerobic conditions is 10-fold lower than that expected from a stoichiometry of one molecule of O2 consumed per molecule of hydrogen peroxide formed attests to the presence of H2O2 scavenging mechanisms in strain D39. However, the presence of other O2 consuming activities, such as NADH oxidase, cannot be ruled out, but the similarity of the carbon and redox balance values observed for strain D39 (Table 2) is in stark disagreement with the presence of high NADH-oxidase activity.

When environmental oxygen is provided continuously (constant partial air tension of 40%), the yield of strain D39 was dramatically reduced to about 30% of the semi-aerobic level, while no major difference in biomass was detected for strain D39spxB (Figure 4 and Figure 8). In terms of fermentative products, oxygenation induced a mild shift to acetate production in the spxB mutant, denoting an alternative aerobic acetate producing pathway. The presence of a functional pyruvate dehydrogenase complex in the pneumococcus has been suggested, but is still to be proven [32,34]. On the other hand, lactate production was severely decreased and a switch to acetate and H2O2, which accumulated up to 2-3 mM, was observed in strain D39. This metabolic shift is in accordance with a higher cell demand for ATP, and is well known to occur in other as a response to stress conditions (aerobiosis, substrate limitation, poor substrates, etc.) [55,56]. Despite the substantially higher yields of acetate in strain D39, the biomass produced was much lower and the ATP concentration in exponential cells was comparable to that under semi-aerobic conditions (data not shown). An important implication from the combined results is that ATP is being redirected to other purposes, seemingly H2O2 detoxification [19]. This view is further corroborated by the growth properties of strain D39spxB under the same conditions.

Despite the modest response of spxB deletion in the transcriptome of strain D39 under semi-aerobic conditions, we show in this study that SpxB affects the pattern of carbohydrate utilization in , and thereby is likely to contribute to the fitness of this microorganism in specific microenvironments. In line, it is reasonable to assume that the influence of environmental oxygen on capsule production and to a lesser extent on growth of is partially concurred by modulation of pyruvate oxidase activity. Interestingly, oxygen is one of the main environmental factors varying in host niches. In this light, the observed connection between capsule amount and oxygen-dependent SpxB activity is plausible. Based on our data we propose that in the nasopharynx, where the oxygen is high, improved SpxB activity restrains capsule production facilitating effective colonization. In contrast, in more internal host niches with no or very limited oxygen, SpxB activity is reduced which relieves the repressing effect on capsule production. This leads ultimately to a thicker capsule that allows to better evade the host immune system.

Materials and Methods

Bacterial strains and growth conditions

Strains used in this study are listed in Table 4 and were stored in glycerol at -80 °C. was routinely grown as standing cultures in M17 broth (Oxoid) [57] containing 0.5% (wt/vol) glucose (Glc-M17) at 37 °C. When necessary erythromycin at a concentration of 0.25 µg ml-1 was used. To detoxify the H2O2 produced, strains were grown in BHI (Oxoid) with or without the addition of 200 U bovine liver catalase (Sigma, St Louis, USA) [58]. BHI medium was used because it lacks antioxidants such as ascorbic acid, which could interfere or obscure effects of the addition of catalase. For growth on plates, 3% (vol/vol) defibrinated sheep blood (Johnny Rottier, Kloosterzande, the Netherlands) was added to Glc-M17 agar.

Table 4

Strains and plasmids used in this study.

Strains	Description	Reference/Source
S. pneumoniae
D39	Serotype 2 strain	[68]
TIGR4	Serotype 4 strain	[69]
D39 pGh9 T7	D39 spxB: :pGh9:ISS1; EmR	This study
TIGR4 pGh9 T7	TIGR4 spxB: :pGh9:ISS1T7; EmR	This study
D39spxB	D39 spxB::ISS1	This study
D39spxB +	D39spxB harbouring pMU1328::300+spxB; EmR	This study
D39pORI Pcps	D39 harbouring pORI Pcps-lacZ; EmR	[33]
D39pORI Pcps spxB	D39 spxB::ISS1 harbouring pORI Pcps-lacZ; EmR	This study
Plasmids
pGh9:ISS1T7	Insertion mutagenesis plasmid with T7 RNA polymerase promoter	[30]
pMU1328::300+spxB	Plasmid carrying the spxB gene under its native promoter; EmR	[19]
pORI13	ori + repA - ; promoterless lacZ; EmR	[70]
pORI Pcps	pORI13 carrying cps promoter in front of the promoterless lacZ ; EmR	[33]

Abbreviations: EmR, erythromycin resistance

Generation of mutants in

D39 was transformed with pGh9 T7 as described before [30]. To excise the plasmid from the genome of , which leaves behind one copy of the ISS1 element [59], the strain containing the pGh9 T7 plasmid was grown overnight at 28 °C after which serial dilutions were plated on Glc-M17 agar without antibiotics. Erythromycin sensitive colonies were screened for further study. The site of insertion into the genome was determined using a PCR based method as described in [60].

To generate a D39spxB ΔccpA mutant the ΔccpA mutation was amplified from chromosomal DNA of strain TIGR4 ΔccpA using primers SP2000 (5' GAACAGTCCGAAACTATGTC 3') and sp1998 (5' ATTCCAGTCGCTCTGGTATC 3') and the resulting PCR product was used for transformation. To generate a spxB strain in which the ΔccpA mutation was ectopically complemented, the chromosomal DNA of the corresponding strain was transformed into the spxB strain [61].

Complementation of the spxB mutation

To generate a complemented D39spxB strain, the pMU1328::300+spxB plasmid, containing the spxB gene under its native promoter and an erythromicin resistance marker [19], kindly provided by Dr. Jeffrey Weiser, was transformed into D39spxB. Erythromycin resistant colonies were selected for further study.

Transmission electron microscopy of

Aliquots of were quickly thawed at 37 °C and diluted 1:40 in Glc-M17 and grown under semi-aerobic conditions at 37 °C until mid exponential phase (OD600 0.2-0.3) Bacteria were harvested by centrifugation for 10 minutes at 3000 x g at 4 °C and subsequently stained with lysine ruthenium red (LRR) and treated as described before [12]. For visualization a Philips EM201 (maximum resolution of ~ 0.5 nm) electron microscope was used.

Determination of transcription using lacZ

To measure transcription of the cps operon, a pORI13 derivative containing the D39 cps promoter was used in the D39 and D39spxB backgrounds [33]. Strains were grown standing in chemically defined medium (CDM) containing 1% Glc (wt/vol) and harvested at mid-exponential (≈ OD600 0.2) or late exponential (≈ OD600 0.4) phase. Specific activity of the cps promoter was assayed by measuring β-galactosidase activity as described before [33].

Transcriptome analysis

For DNA microarray analysis, D39 wild type and two independently obtained D39spxB mutants were grown as four, two and two biological replicates, respectively, in Glc-M17 under semi-aerobic conditions and harvested at an OD595 of approximately 0.25 (mid-exponential). The RNA of both D39spxB strains was compared to D39 and the combined data was analyzed. All other procedures regarding microarray analyses were done as described before [62]. A gene was considered differentially expressed when the fold change was ≥ 2, ≤ 0.5, with a Bayes p ≤ 0.00001 and when at least 7 measurements were available. The microarray data is submitted to the GEO database, GSE45971.

Growth conditions for metabolic analysis

Cells were grown at 37°C in static rubber-stoppered bottles (500-ml) in complex medium M17 (DifcoTM), supplemented with Glc 1% (wt/vol) and without pH control (initial pH 6.5). Prior to inoculation sterile air was bubbled through the medium for 20 min. Alternatively cells were grown in 2-l bioreactors in the CDM described by 46 with the pH controlled at 6.5, and under initial specific air tension of 50-60% (semi-aerobic, maximum dissolved O2 concentration of 130 µM) or aerobic (continuous specific air tension of 40%, constant dissolved oxygen concentration of 90 µM) conditions. In the bioreactors, dissolved oxygen was monitored with a polarographic oxygen electrode (Mettler-Toledo International). The electrode was calibrated to zero or 100% by bubbling sterile argon or air through the medium, respectively. A specific air tension of 40% was maintained by automatic control of the airflow and agitation. Irrespectively of the growth conditions, Glc (1% wt/vol) was used as carbon source. Cultures were inoculated to an initial optical density at 600 nm (OD600) of 0.05 with cells growing exponentially in the same medium. Growth was monitored by measuring the OD600. Specific growth rates (μ) were calculated through linear regressions of the plots of ln(OD600) versus time during the exponential growth phase. A factor of 0.39, determined from a dry weight (DW) (mg ml-1) versus OD600 curve, was used to convert OD600 into dry weight (mg biomass ml-1).

Quantification of fermentation products during growth on glucose

Strains were grown in Glc-M17 or Glc-CDM with pH control (pH 6.5) as described above. Culture samples (2 ml) were taken at different time-points of the growth curves, centrifuged (16,000 × g, 2 min, 4°C), filtered (Millex-GN 0.22 µm filters), and the supernatant solutions were stored at -20°C until analysis by high performance liquid chromatography (HPLC). Substrates and end-products were quantified as before [46] in an HPLC apparatus equipped with a refractive index detector (Shodex RI-101, Showa Denko K.K., Japan) using an HPX-87H anion exchange column (Bio-Rad Laboratories Inc., California, USA) at 60°C, with 5 mM H2SO4 as the elution fluid and a flow rate of 0.5 ml min-1. Alternatively, quantification of metabolites in supernatants was performed by 1H-NMR in a Bruker AMX300 spectrometer (Bruker BioSpin GmbH). Formic acid (sodium salt) was added to the samples and used as an internal concentration standard. The ATP yield was calculated as the ratio of ATP produced to glucose consumed. The global yields of ATP were calculated from the fermentation products determined at the time-point of growth arrest assuming that all ATP was synthesized by substrate-level phosphorylation.

Determination of hydrogen peroxide (H2O2)

Strains were grown in Glc-CDM as described above. Culture samples of 1-ml were harvested at different time-points of the growth curves, centrifuged (16,000 × g, 2 min, 4°C) and filtered (Millex-GN 0.22 µm filters). The Amplex® Red Hydrogen Peroxide/Peroxidase assay kit (Invitrogen) was used to quantify hydrogen peroxide contents below 10 µM. Determinations of H2O2 up to 300 µM were performed as described elsewhere [63]. Briefly, 1.25 ml of peroxide reagent (192 mM phosphate, 14.8 mM azide, 0.96 ml l-1 Triton X-100, 2 KU l-1 horseradish peroxidase (Roche), 0.48 mM 4-aminophenazone and 9.6 mM chromotropic acid) was added to 50 µl of supernatant, mixed and allowed to stand for 5 minutes at room temperature. In the presence of H2O2, the chromotropic acid was converted by the peroxidase into a blue coloured compound with maximal absorbance at 600 nm. Absorbance was read at 600 nm. Water was used as the blank and standard curves were performed with fresh dilutions of a stabilized solution of 30% (wt/vol) H2O2. The absorbance of the samples was compared to that of the standard solutions.

Determination of intracellular metabolites during growth on glucose

Strains were grown in Glc-CDM under semi-aerobic as described above. Ethanol extracts for analysis by 31P-NMR and quantification of phosphorylated metabolites in D39spxB and control strain D39 at late-exponential (LExp) and early-stationary (EStat) phases of growth (time-points indicated by the arrows in Figure 4B) were prepared as previously described by 46. In brief, a volume of culture was withdrawn from the fermenter, adjusted to obtain approximately 16-18 mg of cell protein for extract, and centrifuged (21,000 x g, 15 min, 4°C). The pellet was suspended in 28 ml of 5 mM 3-(N-morpholino) propanesulfonic acid (MOPS) buffer, pH 7.2 (Sigma-Aldrich) transferred to 70 ml of cold ethanol (final concentration 70% vol/vol) in an ice bath, and extraction was performed for 30 min with vigorous agitation. Cell debris was removed by centrifugation (35,000 x g, 30 min, 4°C), the ethanol was evaporated and the residue lyophilized. The dried extracts were dissolved in 1 ml 2H 2O containing 5 mM EDTA (final pH approximately 6.5) and stored at -20°C. Assignment of resonances was based on previous studies [46] or by spiking the sample extracts with the suspected, pure compounds. The concentrations of intracellular metabolites were calculated from the areas of their resonances in 31P-spectra by comparison with the area of the resonance due to methylphosphonic acid (Sigma-Aldrich), added as an internal standard, and after application of an appropriate factor for correcting saturation of resonances. The reported values for intracellular phosphorylated compounds are averages of three to four independent growth experiments. The accuracy was around 15-20%. Phosphorus-31 spectra were acquired on a Bruker AVANCE II 500 spectrometer (Bruker BioSpin GmbH) with a phosphorus selective 5-mm-diameter probe head at 30°C with a pulse width of 6.5 µs (flip angle, 70°) and a recycle delay of 1s. Saturation of resonances due to fast pulsing conditions was calculated by comparison with spectra acquired under fully relaxed conditions (recycle delay, 20s). Phosphorus chemical shifts were referenced to the resonances of external 85% H3PO4 designated at 0.0 ppm.

Capsular polysaccharide preparation and quantification of the glucuronic acid content

Samples for the determination of the capsular glucuronic acid amounts were prepared as follows: cells grown in Glc-CDM under semi-aerobic conditions as above and harvested during the late-exponential and early-stationary phases of growth (time-points indicated by the arrows in Figure 4B) were centrifuged (6,000 x g, 3-7 min, 4°C), resuspended in PBS and pelleted at 3000 x g, 4°C, for 20 min. The pellet was resuspended in 500 µl of 150 mM Tris-HCl (pH 7.0) and 1 mM MgSO4 and treated as described elsewhere [64]. The supernatants derived from the two centrifugations referred before were pooled and treated with 20% (wt/vol) trichloroacetic acid for protein precipitation. After 2 h of incubation in an ice bath proteins were pelleted and the exopolysaccharides were precipitated with cold ethanol as in [65]. The glucuronic acid of capsule attached or loosely attached (lost in centrifugations to the supernatants) to the cell wall was quantified by the method for quantitative determination of uronic acids as described in [66]. Capsule measurements were performed in duplicate using samples from six independent cultures.

Growth of D39spxB and D39 strains on specified sugars

To test the ability of specific sugars to support growth of strains D39 and D39spxB cultures of 250 µl were prepared in CDM containing 0.25% (wt/vol) of glucose (Glc), glucuronic acid (GlcUA), rhamnose (Rha), gluconic acid (GlcA), N-acetylglucosamine (GlcNAc) or a 1:2:3 mixture of GlcUA, Glc and Rha, respectively. Cultures were started at an initial OD595 of 0.25-0.3 by addition of a preculture, grown on Glc-M17 (0.5% Glc wt/vol), suspended in fresh CDM without sugar. Cultures were grown for 24 h at 37°C in 96-well microtiter plates. Cell density was monitored at 595 nm with a ELx808 microplate spectrophotometer (BioTek Instruments, Inc.) and the growth curves were generated using Gen5TM (BioTek Instruments, Inc.) with readings taken every 30 min. Each growth condition was done in triplicate using two independent precultures. Growth rates and other growth parameters were generated as above and the results were analyzed using the Instat software (GraphPad Software).

Detection of SpxB in D39, D39spxB and D39spxB .

Detection of SpxB by Western blot using a ployclonal anti-SpxB serum. D39 (lane 1), D39spxB (lane 2) and D39spxB (complemented strain) (lane 3). Strains were grown in BHI to an OD of 0.2-0.25.

(TIF)

Click here for additional data file.

Supplemental materials and methods.

(PDF)

Click here for additional data file.