The Effects of Methionine Acquisition and Synthesis on Streptococcus Pneumoniae Growth and Virulence

Bacterial pathogens need to acquire nutrients from the host, but for many nutrients their importance during infection remain poorly understood. We have investigated the importance of methionine acquisition and synthesis for Streptococcus pneumoniae growth and virulence using strains with gene deletions affecting a putative methionine ABC transporter lipoprotein (Sp_0149, metQ) and/or methionine biosynthesis enzymes (Sp_0585 - Sp_0586, metE and metF). Immunoblot analysis confirmed MetQ was a lipoprotein and present in all S. pneumoniae strains investigated. However, vaccination with MetQ did not prevent fatal S. pneumoniae infection in mice despite stimulating a strong specific IgG response. Tryptophan fluorescence spectroscopy and isothermal titration calorimetry demonstrated that MetQ has both a high affinity and specificity for L-methionine with a KD of ∼25 nM, and a ΔmetQ strain had reduced uptake of C14-methionine. Growth of the ΔmetQ/ΔmetEF strain was greatly impaired in chemically defined medium containing low concentrations of methionine and in blood but was partially restored by addition of high concentrations of exogenous methionine. Mixed infection models showed no attenuation of the ΔmetQ, ΔmetEF and ΔmetQ/ΔmetEF strains in their ability to colonise the mouse nasopharnyx. In a mouse model of systemic infection although significant infection was established in all mice, there were reduced spleen bacterial CFU after infection with the ΔmetQ/ΔmetEF strain compared to the wild-type strain. These data demonstrate that Sp_0149 encodes a high affinity methionine ABC transporter lipoprotein and that Sp_0585 – Sp_0586 are likely to be required for methionine synthesis. Although Sp_0149 and Sp_0585-Sp_0586 make a contribution towards full virulence, neither was essential for S. pneumoniae survival during infection.


Introduction
The acquisition of essential nutrients from the host is a prerequisite for bacterial pathogens to be able to replicate and so to cause successful infection. Screens for virulence genes as well as targeted investigation of specific nutrient transporters has confirmed the importance of nutrient acquisition for the pathogenesis of infections for many microbial pathogens [1,2,3,4,5,6,7]. One group of nutrients required for bacterial growth are the amino acids, but there are only limited data on their importance for bacterial pathogenesis.
Methionine is one of the least abundant amino acids in physiological fluids (4 mg ml 21 ) [8], and yet is essential for protein synthesis and is a constituent of S-adenosylmethionine, the major biological methyl donor required for the biosynthesis of phospholipids and nucleic acids [9]. Two bacterial methionine transport systems have been described: the Methionine ABC (ATP binding cassette) Uptake Transporter (MUT) family [10,11] and a secondary transporter termed BcaP [12]. In Escherichia coli, the MUT system is encoded by the metD locus and consists of the MetQ substrate binding protein (SBP), MetL transmembrane permease and the MetN cytoplasmic ATP-hydrolyzing protein (ATPase) [11]. E. coli metD mutants are unable to transport Dmethionine or utilize this compound as a source of methionine [11]. Similar ABC transporters are the primary methionine transporters for Bacillus subtilis (metNPQ) [10], Streptococcus agalactiae (metQ1NP) [13] and Streptococcus mutans (atmBDE) [14]. The second methionine transport system, BcaP was described in Lactococcus lactis as a branched chain amino acids transporter, but is also involved in the transport of methionine [12]. Microorganisms and plants can also synthesize methionine by converting homoserine to homocysteine through addition of a sulphur group from either cysteine (requiring MetABC), sulfide (requiring MetA and CysD) or methionine using the SAM recycling pathway (MetK, Pfs, and LuxS) [15,16]. Homocysteine is then methylated by methionine synthase (MetE) in conjunction with a methylenetetrahydrofolate reductase (MetF), with the methyl group supplied by 5-methyl tetrahydrofolate, to form methionine [15,17]. Existing data show that methionine biosynthetic genes are required for the full virulence of Brucella melitensis [18], Haemophilus parasuis [19] and Salmonella enterica [20] and that mutation of the S. agalactiae methionine regulator MtaR attenuates virulence [8], suggesting methionine synthesis is essential for survival of many bacteria during invasive infection.
Streptococcus pneumoniae is a common nasopharyngeal commensal that is also an important pathogen frequently causing pneumonia, otitis media, septicaemia and meningitis. A recent investigation of the role of nine different ABC transporters for S. pneumoniae virulence identified an ABC transporter encoded by Sp_0148-52 that seemed to be important during pneumonia and septicaemia [7]. BLAST searches suggested this locus contained genes whose products have a high degree of identity to MetQNP and AtmBDE and therefore could be a S. pneumoniae methionine uptake ABC transporter. In this manuscript we describe in detail the Sp_0148-52 locus and the role of methionine during S. pneumoniae growth and virulence. Recombinant Sp_0149 was used to characterise the potential substrates of this ABC transporter, and deletion mutant strains of Sp_0149 and metEF were used to investigate role of methionine acquisition and synthesis during S. pneumoniae growth and virulence. In addition, as several S. pneumoniae lipoproteins have been shown to be effective vaccine candidates in animal models [21,22], we also investigated the potential of recombinant Sp_0149 as a novel vaccine candidate.

Results of BLAST alignments for Sp_0148-52
The Sp_0148-52 genetic locus was identified during a screen of ABC transporters for those involved in virulence. In this screen a mutant containing an insertion within Sp_0149 was attenuated in virulence in mouse models of pneumonia and sepsis [7]. The Sp_0148-52 region in the TIGR4 S. pneumoniae strain genome contains five genes and is 3669 bp in length ( Fig. 1 A) [23]. Gene loci with identical organisation and high levels of identity at the amino acid level for the corresponding genes are present in S. agalactiae (SAN_1757-52) [13] and S. mutans (Smu.1942-38) [14] and contain the metQNP and atmBDE genes respectively that encode probable methionine ABC transporters. Both Sp_0148 and Sp_0149 possess potential lipoprotein signal sequences, LAACS (residues [21][22][23][24][25] and LAACG (residues 20-24) respectively, and are annotated in the TIGR4 genome as lipoproteins. Sp_0148 and Sp_0149 only share 21% identity and 34% similarity at the amino acid level, so are unlikely to have similar functions or have arisen by gene duplication. Sp_0148 has been annotated as an amino acid binding lipoprotein and is highly conserved amongst sequenced S. pneumoniae strains (98% identity at the amino acid level) and with proteins encoded by other streptococci (.49% identity). Sp_0149 is conserved in all the available S. pneumoniae genomes (.97% identity at the amino acid level) and has 62% identity at the amino acid level to the lipoprotein component of the S. mutans methionine transporter atmBCDE [14], suggesting it may have a similar function. BLAST searches demonstrate 73% identity at the amino acid level between the predicted proteins encoded by Sp_0151, atmD and metN (all predicted ATPases) and 70% for the predicted proteins encoded by Sp_0152, atmE and metP (all predicted permeases). Hence, Sp_0149-52 could encode a functional S. pneumoniae methionine ABC transporter, and therefore we have named Sp_0149 metQ, Sp_0151 metN, and Sp_0152 metP according to the nomenclature for the genes encoding a putative S. agalactiae methionine ABC transporter [13].
To investigate whether S. pneumoniae metQNP encode a methionine ABC transporter and the phenotype attributable to loss of its function, a metQ deletion mutant strain of 0100993 S. pneumoniae was constructed using a construct made by overlap extension PCR (Fig. 1A). Complete deletion of Sp_0149-0152 was not done as this would include deletion of Sp_0150 as well as metQNP. BLAST results suggested Sp_0150 encodes a protein belonging to the M20/M40/M45 family of peptidases. Genes encoding similar proteins are found within the S. mutans (atmC) and S. agalactiae (pdsM) methionine ABC transporter loci, but the function of the corresponding proteins is unknown. In addition the Sp_0148 lipoprotein may also utilise the MetN and MetP proteins to form an ABC transporter, making it difficult to attribute any phenotypes of an Sp_0149-0152 deletion mutant specifically to a single function. The correct identity of the mutant was confirmed using PCR and primers annealing to Sp_0148 or Sp_0150 and the erythromycin resistance gene ery r (data not shown). Attempts at complementing the metQ deletion mutant strain using reintegration of an intact copy of the full length gene were unsuccessful. However, qPCR demonstrated that expression of the downstream genes in the operon was not reduced by the DmetQ mutation, with fold differences (SD) in transcript levels from the DmetQ strain compared to the wild-type strain of 2.2 (2.3), 1.3 (1.2), and 1.1 (0.2) for Sp_0150, Sp_0151, and Sp_0152 respectively.

Localisation studies of Sp_0149 lipoprotein
To confirm whether the lipoprotein MetQ is localised to the S. pneumoniae cell membrane, polyclonal anti-MetQ antibodies were obtained by intraperitoneal inoculation of recombinant MetQ into mice. Serum obtained after 4 weeks from inoculated mice recognised recombinant MetQ protein ( Fig. 2A and B), and immunoblots of whole cell wild-type S. pneumoniae lysates identified a single band of similar size to the predicted size of MetQ excluding the lipoprotein signal peptide sequence. The same band was present when Triton X-114 extracts of membrane bound proteins were probed with polyclonal anti-MetQ antibodies but was only present as a weak band in the aqueous fraction ( Fig. 2A). Previous studies have demonstrated that cleavage of the N terminal lipoprotein signal peptide sequence of lipoproteins requires lipoprotein signal peptidase (Lsp), and that lipoproteins in the S. pneumoniae Dlsp mutant are therefore slightly larger than in the wild type strain [24]. Immunoblotting of whole cell lysates and aqueous fractions of the Dlsp S. pneumoniae strain with anti-MetQ serum demonstrated that, as predicted, the band representing MetQ was slightly larger than in the wild-type strain ( Fig. 2A). The band representing MetQ in the aqueous extracts from the Dlsp strain was also more prominent that in the wild-type strain, perhaps suggesting reduced retention of MetQ in the cell membrane in the absence of the Lsp enzyme. Overall, the immunoblot results demonstrate that MetQ is mainly associated with the membrane fraction and is processed by Lsp, confirming that MetQ is a lipoprotein.
MetQ binds L-methionine with high affinity and is required for methionine uptake by S. pneumoniae Potential substrates for the MetQNP ABC transporter were investigated using fluorescence spectroscopy of recombinant MetQ (Supplementary Figure 1), a method widely used to investigate ligand specificity for SBPs due to the large conformational changes that occurs upon ligand binding [25,26]. We observed an approximately 30% enhancement of the intrinsic fluorescence of the protein in the presence of L-methionine ( Supplementary Fig. 2) and we determined a K D of MetQ for L-methionine to be determined as 26.5861.15 nM (Fig. 3A). We confirmed this tight binding using isothermal titration calorimetry, which yielded almost identical binding parameters of 26.5561.28 nM (Fig. 3B), and also demonstrates 1:1 binding of L-methionine to MetQ.
To investigate the specificity of MetQ for L-methionine, we examined ligand binding using a range of related ligands. Binding of D-methionine was around 1000 fold weaker with a K D of 27.460.96 mM, while binding of DL-homocysteine was only around 250-fold weaker with a K D of 7.060.26 mM (Supplementary Fig. 3). We also detected weak binding of a-methyl-DLmethionine with a K D of 30.160.86 mM ( Supplementary Fig. 3), but could not detect any binding for either L-cysteine of glycine ( Supplementary Fig. 2). These data demonstrate that MetQ has a high affinity and specificity for L-methionine, and confirm that methionine is the probable major ligand for the S. pneumoniae MetQNP ABC transporter. To confirm that MetQ is required for methionine uptake by S. pneumoniae, uptake assays of 14 C-labelled methionine were performed in chemically defined medium with the wild-type and DmetQ strains [27]. As the mucoid phenotype of the serotype 3 strain prevents the efficient formation of bacterial pellets necessary for these assays, the DmetQ mutation was transferred into a serotype 2 S. pneumoniae strain (D39). In C+Y medium without added methionine, the DmetQ mutant strain was unable take up C 14 -methionine whereas the wild-type strain demonstrated significant uptake (Fig. 3C). In the presence of 750 mM methionine there was uptake of C 14 -methionine by the DmetQ mutant strain suggesting the presence of a lower affinity methionine transport mechanism (Fig. 3C). These results confirm that metQNP is a methionine ABC transporter that is required for methionine uptake by S. pneumoniae when methionine availability is limited.

Investigation of the lipoprotein MetQ as a vaccine candidate
Lipoprotein are attached to the external surface of bacterial membranes and have been shown to be effective vaccines that protect against S. pneumoniae infection in mouse models [21,22]. Hence, MetQ was investigated as a potential novel vaccine candidate. PCR (data not shown) and immunoblots demonstrated that MetQ was present in all of a range of S. pneumoniae strains from common MLST sequence types for clinically important capsular serotypes (Fig. 2B). Mice were immunised by three intraperitoneal injections of 10 ug recombinant MetQ with alum as the adjuvant separated by 7 or 8 days, which resulted in significant serum titres of IgG (mainly IgG1) against MetQ that did not cross-react with the control lipoprotein LivJ (2) (Fig. 4A). However, when challenged by intraperitoneal inoculation with a potentially lethal dose of the serotype 2 S. pneumoniae strain D39 there were no differences in the speed of development of fatal infection and survival between vaccinated and control (given PBS and the adjuvant alone) mice (Fig. 4B). To investigate whether this lack of vaccine efficacy despite good anti-MetQ titres was due to lack of MetQ expression by S. pneumoniae during infection, semiquantitative reverse transcriptase PCR (RT-PCR) for metQ transcripts was performed using S. pneumoniae RNA extracted from the blood of mice with severe S. pneumoniae septicaemia. As a positive control RT-PCR was also performed for psaA, which encodes another lipoprotein that is expressed during infection. RT-PCR readily identified a product of the right size for metQ, but at a lower level of expression compared to psaA, suggesting that relative lack of expression of metQ could be one reason why vaccination with the MetQ was not protective (Fig. 5B). These data were supported by transcriptome data calculating the relative abundance of metQ, metE, metF, PsaA and another lipoprotein PiuA at three anatomical sites during infection (Fig. 5C). The results show that transcript abundance for metQ, metE and metF were similar for bacteria found in the nasopharynx, lungs and blood. Importantly, the amount of metQ transcripts was almost two log 10 lower and one log 10 lower than the amount of psaA and piuA transcripts respectively, both of which encode effective lipoprotein vaccine candidates [28,29,30,31]. These results support the hypothesis that lack of MetQ adundance during infection could explain the failure of vaccination with MetQ to provide protection against systemic disease. Furthermore, when assessed using an established flow cytometry assay polyclonal anti-MetQ sera did not significantly improve phagocytosis of other S. pneumoniae strains by a neutrophil cell line (Fig. 4C).
In vitro phenotypes of the DmetQ, DmetEF and DmetEF/ DmetQ strains As synthesis of methionine may reduce the effects of loss of transport of methionine due to the metQ mutation on S. pneumoniae phenotype, additional mutant strains were constructed containing deletions of the S. pneumoniae metE and metF [32]. BLAST searches of S. pneumoniae TIGR4 genome identified the adjacent (and likely cotranscribed) genes Sp_0585 and Sp_0586 as homologues of the S. mutans metE and metF genes (47% and 38% identity at the amino acid level respectively) [14] (Fig. 1B). The genome structure suggested metE and metF were likely to be contranscribed independent of the adjacent genes, due to the large gap between Sp_0584 and Sp_0585, and because Sp_0587 is transcribed in the opposite direction to Sp_0586 (Fig. 1B). Deletion mutants replacing metE and metF with a chloramphenicol resistance marker were made in the wild-type (DmetEF) or in combination with metQ deletion (DmetEF/DmetQ) S. pneumoniae 0100993 strains (Fig. 1B). The correct identity of the mutants was confirmed using PCR and primers annealing to Sp_0584 or Sp_0587 and the chloramphenicol resistance gene cat (data not shown).
To investigate the functional importance of methionine synthesis and uptake for S. pneumoniae, growth of the wild-type, DmetQ, DmetEF and DmetEF/DmetQ strains were compared in the complete medium THY and in the semi-defined medium C+Y with restricted methionine content. In THY there were no significant differences in growth between the wild-type and DmetQ, DmetEF and DmetEF/DmetQ strains (Fig. 6A). In C+Y without added methionine (Fig. 6B), both DmetQ and DmetEF mutant strains had slightly impaired growth compared to the wild-type strain, but still reached similar growth levels to the wild-type strain after 9 hours. However, in C+Y with either no added methionine or supplemented with only 50 mM of methionine the double mutant DmetEF/DmetQ had markedly impaired growth due to reduced generation times ( Fig. 6B and C). These results provide further evidence that MetQ is required for methionine uptake and indicate that MetEF are necessary for methionine synthesis. The impaired growth of the DmetEF/DmetQ strain in C+Y medium was partially restored by addition of 750 mM of methionine (Fig. 6D), suggesting there are additional weaker affinity secondary transport systems that allow the DmetEF/DmetQ strain to obtain methionine in the presence of high concentrations of substrate.
To assess the requirement for methionine for S. pneumoniae growth in physiologically relevant conditions, replication of these strains was assessed in freshly obtained human blood. The DmetQ and DmetEF strains were both significantly impaired in growth in blood (inoculum size 1.5610 6 CFU, 10.9 and 7.7 fold increase in CFU, respectively) compared to the wild type strain (inoculum size 0.5610 6 CFU, 16.2 fold increase, P,0.001, ANOVA) (Fig. 7A), although all three strains reached similar total CFU per ml. Moreover, the DmetEF/DmetQ strain had markedly reduced growth in blood with only 3.8 fold increase in CFU after 4 hours and reaching a total CFU of nearly 1 log 10 fewer than the wild-type strain (Fig. 7A). Addition of 1 mM methionine but not leucine (a branched-chain amino acid not linked to the methionine biosynthetic pathways) largely reversed the growth defect of the double mutant in blood, without affecting growth of the wild-type strain (Fig. 7B). These results demonstrate that absence of methionine uptake and synthesis strongly impairs S. pneumoniae growth in physiological conditions found during infection.
In vivo phenotypes of the DmetQ, DmetEF and DmetEF/ DmetQ strains To investigate the effects of deletion of metQ and/or metEF on the ability of S. pneumoniae to colonise the nasopharynx, competitive indices (CIs) for mixed infection with the wild-type and the DmetQ, or DmetEF or DmetEF/DmetQ strains were determined in a mouse model of colonisation. None of the mutant strains were reduced in virulence in the colonisation model when CIs were calculated for nasal washes obtained 5 days after inoculation (Fig. 8A). Previously we have shown that a metQ disruption strain was attenuated in virulence in mouse models of sepsis and pneumonia when investigated using CIs [7]. To further assess whether the DmetQ mutation affected the development of infection, groups of 5 CD1 mice each were intranasally inoculated with 10 7 CFU of the wildtype or DmetQ strain, and bacterial CFU from the BALF, lungs and spleens determined after 48 hours (Fig. 8B). There were no significant differences in the bacterial loads between the wild-type and DmetQ strains in the BALF, lungs and spleens. Hence although previously we have shown disruption of metQ strain reduced virulence in a competitive model of lung infection [7], the DmetQ strain was still able to cause significant infection when given as a pure inoculum. The severe growth defect of the double mutant DmetEF/DmetQ in human blood prompted us to investigate its growth in a mouse sepsis model. After intraperitoneal inoculation the number of DmetEF/DmetQ strain CFU recovered from the blood was approximately two log 10 fewer than CFU recovered for the wild-type strain (P = 0.032, Mann Whitney U test) (Fig. 8C). These data suggested that the inability of the DmetEF/DmetQ strain to either transport or synthesise methionine has resulted in decreased but not completely abrogated virulence of this strain in a mouse sepsis model.

Discussion
In this study, we provide evidence that the S. pneumoniae Sp_0148-52 locus encodes a methionine uptake ABC transporter and that Sp_0585 and Sp_0586 encode MetE and MetF required for methionine biosynthesis. BLAST searches demonstrated that the Sp_0148-52 and Sp_0585/Sp_0586 encode proteins with high degrees of similarity to S. mutans AtmBDE and MetEF respectively, which have been shown to be required for methionine uptake and synthesis [13,14]. Deletion of the gene encoding the lipoprotein component Sp_0149 (which we have called metQ) resulted in a strain that had reduced growth in methionine restricted media and no detectable uptake of radioactive methionine. Tryptophan fluorescence spectroscopy confirmed MetQ bound to L-methionine with high affinity as well as D-methionine and homocysteine, an amino acid that is related to methionine. Deletion of Sp_0585/ Sp_0586 increased the growth defect of the metQ deletion strain in  methionine restricted media and in blood, supporting a role for the products of these genes for methionine synthesis. Genetic complementation would be helpful to link the metEF and metQ mutations to these phenotypes but was not possible to achieve, as is often the case with S. pneumoniae. However the successful biochemical complementation of the in vitro phenotypes by exogenous methionine and the lack of impaired transcription of the rest of the operon in the metQ mutant strongly support that the observed phenotypes were caused by the mutations. Overall the data indicate that S. pneumoniae methionine acquisition and synthesis is largely dependent on the Sp_0148-52 locus and Sp_0585 and Sp_0586 loci respectively. Unlike the equivalent S. mutans mutants [14], the S. pneumoniae metQ and metEF mutant strains were still able to grow in the absence of supplementation with methionine. This observation and the persisting growth of the S. pneumoniae DmetQ/metEF double mutant in media containing high concentrations of methionine suggest that there are additional mechanisms for S. pneumoniae methionine uptake, possibly similar to BcaP of L. lactis [12]. One candidate is the Sp_0481-4 locus which contains an upstream consensus sequence for the streptococcal methionine regulator MtaR/MetR (data not shown).
As well as encoding the MetQNP methionine ABC transporter, the Sp_0148-52 locus also encodes another SBP (Sp_0148) unrelated to MetQ and a peptidase (Sp_0150). The functions of these proteins and why their genes are associated with the metQNP genes are unknown. The peptidase could be required for the breakdown of extracellular peptides to provide substrates for MetQNP, but this hypothesis will require further investigation. Previously it has been suggested that two separate zinc uptake lipoproteins AdcA and AscAII share the same permease and ATPase components [33]. Similarly, the Sp_0148 SBP could associate with MetN and MetP to make an ABC transporter with separate affinities to the MetQ lipoprotein, perhaps another amino acid as Sp_0148 has some similarities to proteins annotated as glutamine SBPs.
Existing vaccines against S. pneumoniae are unable to provide protection against all strains, prompting the investigation of surface exposed antigens such as SBPs that are generally highly conserved between different S. pneumoniae strains as novel vaccine candidates. Existing data show that several ABC transporters SBPs are effective vaccines in animal models of S. pneumoniae [28,30,34,35]. However, vaccination with MetQ did not prevent fatal systemic infection despite inducing a reasonable antibody response and the surface location of MetQ. RT-PCR and transcriptome data suggested that expression of metQ during actual infection was considerably lower than that of the SBP vaccine candidates psaA and piuA [36]; this low level of expression could readily contribute to the lack of protective efficacy of high titres to the S. pneumoniae MetQ. In addition, unlike IgG raised against other SBPs [31,34,35], anti-MetQ antibodies did not promote neutrophil phagocytosis of S. pneumoniae, suggesting these antibodies are functionally ineffective at promoting immunity.
Several S. pneumoniae ABC transporters have been shown to be required for full virulence, including those involved in the acquisition of metals (manganese, zinc, and iron), phosphate, polyamines, and branched chain amino acids [5,6,7,29,34,35,37]. As methionine is essential for bacterial growth and has restricted availability in mammalian systems, with a concentration of 4 mg ml 21 in physiological fluid, the MetQNP ABC transporter might also be expected to be important for S. pneumoniae virulence. Indeed, our previous data using mixed infection experiments with the wild-type strain and a metQ mutant created by insertional mutagenesis showed the metQ mutant strain was significantly less competitive in both pneumonia and sepsis infection models [7]. However, the competitive nature of the mixed infection model makes it a very stringent test of virulence for mutant strains, and in this manuscript we have shown that when given a pure inoculum of the deletion DmetQ or wild-type strains the progression of infection in mouse model of pneumonia was very similar. These results are similar to those we have described for the S. pneumoniae branched chain amino acid ABC transporter, which also has reduced virulence when competing against the wild-type strain but could still establish fatal infection [7]. The double mutant DmetEF/metQ strain was attenuated in virulence in a sepsis model with lower bacterial CFU than the wild-type strain, suggesting that methionine synthesis helped to maintain virulence of the DmetQ strain. However, somewhat surprisingly the mutant DmetEF/metQ strain was still able to establish systemic infection in mice, despite its poor growth in blood in vitro and a circulating methi-onine level in mammalian blood of 7 to 43 mM, a level that is associated with a strong growth defect for the this strain in medium. Similarly, there was no evidence that deletion of either metQ or metEF individually or in combination affected colonisation when assessed using highly sensitive competitive infection assays. These results may indicate that local methionine availability at the site of infection is relatively high, perhaps due to cell lysis, and adequate for S. pneumoniae growth even in a strain containing deletions affecting both high affinity methionine uptake and methionine biosynthesis. The effect on virulence of some S. pneumoniae mutations vary with strain background, and it is possible that in other strains metQ and metEF deletion will impair virulence to a greater extent than seen in the serotype 3 strain used for these studies.
To conclude, we have assessed the role of the MetQNP ABC transporter and MetEF methionine synthesis enzymes for S. pneumoniae growth and virulence. The results demonstrate that MetQNP is required for methionine uptake, and that dual mutation of metQ and the metEF locus has a profound effect on growth of S. pneumoniae in methionine restricted conditions. Methionine uptake and synthesis assisted, but was not essential, for the development of S. pneumoniae infections. These data provide more information on the micronutrient requirements of S. pneumoniae during the development of infection.

Ethics statement
All animal experiments were approved by the UCL Biological Services Ethical Committee and the UK Home Office (Project Licence PPL70/6510). Experiments were performed according to UK national guidelines for animal use and care, under UK Home Office licence.

Bacterial strains, media and growth conditions
The E. coli strain M15 (Qiagen) were used for the cloning of Sp_0149 lipoprotein gene. Capsular serotypes 2 D39 strain and capsular serotype 3 S. pneumoniae 0100993 strain were used to construct S. pneumoniae mutant strains for the in vitro and in vivo phenotype analysis [1]. A range of S. pneumoniae strains from common multi-locus sequence typing (MLST) sequence types (STs) for clinically important capsular serotypes were kind gifts from Professor Spratt, Imperial College [38]. E. coli was cultured at 37uC using Luria Bertani (LB) broth or agar plates and S. pneumoniae strains were cultured in the presence of 5% CO 2 at 37uC on Columbia agar (Oxoid) supplemented with 5% horse blood (TCS Biosciences), or in Todd-Hewitt broth supplemented with 0.5% yeast-extract (Oxoid) or C+Y media [39]. Plasmids and mutant strains were selected by using appropriate antibiotics (carbenicillin 100 mg ml 21 and kanamycin 25 mg ml 21 for E. coli, 0.2 mg ml 21 erythromycin for S. pneumoniae). Stocks of S. pneumoniae were stored as single use 0.5 ml aliquots of THY broth culture (O.D 580 0.3-0.4) at 270uC in 10% glycerol. Growth of S. pneumoniae strains in broth was monitored by measuring optical density at 580 nm (OD 580 ). Replication of S. pneumoniae strains in freshly obtained human blood was determined by inoculating with 0.5 to 2.0610 6 CFU ml 21 (depending on strain and experiment). After 2, 4, 6 h of growth at 376C under 5% CO 2 , serial dilutions were plated on to blood agar plates to enumerate the CFU. Fold change was calculated by dividing the results by the inoculum CFU.
Nucleic acid manipulations, reverse transcriptase PCR and transcriptome analysis S. pneumoniae genomic DNA was extracted from bacteria grown in THY using a modified Wizard genomic DNA kit (Promega) and RNA using SV total RNA extraction kit (Promega) as previously described [6]. Total RNA from S. pneumoniae found in mouse blood was extracted as described by Ogunniyi et al [40]. S. pneumoniae were harvested from human and mouse blood by brief centrifugation at 825 g at 4uC for 5 minutes. The resulting supernatant was centrifuged further at 15500 g at 4uC for 5 minutes, and the bacterial pellet resuspended in 300 ml of pre-warmed (65uC) acidphenol (Ambion) and incubated for 5 minutes at 65uC followed by further addition of 300 ml of pre-warmed NAES buffer and incubation at 65uC for 5 minutes with intermittent mixing. The mixture was cooled on ice for 1 minute and centrifuged at 15500 g for 1 minute, and the aqueous phase was re-extracted twice with acid-phenol and NAES buffer followed by twice with 300 ml of chloroform. To this mixture, sodium acetate was added at a final concentration of 300 mM followed by addition of 2 volumes of ethanol and RNA was precipitated at 220uC overnight, centrifuged at 6000 rpm for 5 minutes and washed in 70% ethanol before resuspension in 50 ml of nuclease-free water. To the resulting RNA, recombinant RNasin ribonuclease inhibitor (Promega N251A) was added to a final concentration of 1 U ml 21 and then treated with 0?5 U ml 21 RQ1 RNase-free DNase (Promega M610A) at 37uC for 40 minutes. Aliquots of this RNA were stored at 270uC until use. An aliquot was used to check the purity by RT-PCR with and without reverse transcriptase using gene specific primers. qPCR was performed as previously described [24] using cDNA amplified from 0.4 mg of total RNA using the Applied Biosystems Geneamp RNA PCR core kit (Applied Biosystems, UK), and target-specific primers used for Sp_0150, Sp_0151 and Sp_0152 expression (Table 1). For each gene, crossing point (Cp) values were determined from the linear region of the amplification plot and normalized by subtraction of the Cp value for 16S RNA generating a DCp value. Relative change was determined by subtraction of the DCp value for the wild-type strain from the DCp value for the mutant strain (DDCp value), and fold change calculated using the formula 2 2DDCp . Transcriptome data for RNA extracted from S. pneumoniae serotype 4 strain WCH43 recovered from mouse nasopharynx, lungs and blood was obtained as described [41], using whole-genome S. pneumoniae PCR arrays obtained from the Bacterial Microarray Group at St George's Hospital Medical School, London, England (http://bugs.sghms.ac.uk/). Transcript abundance for each gene was obtained, and quantitative fold differences determined as a ratio of a specific gene abundance to that of the 16S rRNA control.
Restriction digests, ligation of DNA fragments, fractionation of DNA fragments by electrophoresis, and transformation of E. coli (by heat shock) were performed according to established protocols [42]. DNA fragments were purified from electrophoresis gels using the QIAquick gel extraction kit. RT-PCR was performed using the Access RT-PCR system (Promega) and gene specific primers ( Table 1). The NCBI website (http://www.ncbi.nlm.nih.gov/blast) was used to perform BLAST searches and alignments of the available complete and incomplete bacterial nucleotide and protein databases. Primers for both PCR and RT-PCR were designed using sequences displayed by the ARTEMIS software.
Construction of DmetQ, DmetEF and DmetEF/DmetQ S. pneumoniae mutant strains Strains, plasmids and primers used for this study are described in Table 1. Target genes were amplified by PCR from S. pneumoniae 0100993 genomic DNA using primers designed from the TIGR4 genome sequence (http://www.tigr.org) [23]. The DmetQ strain was constructed by overlap extension PCR [43] using a transformation fragment in which the Sp_0149/metQ gene had been replaced by the erythromycin resistance cassette ery. Two products corresponding to 630 bp 59 (primers Sp0148F and Ery-Sp0148R) and 672 bp 39 (primers Ery-Sp0150F and Sp0150R) to metQ were amplified from S. pneumoniae genomic DNA by PCR carrying 39 and 59 linkers complementary to the 59 and 39 portion of the ery gene respectively. ery was amplified from pACH74 using PCR and the primers EryF and EryR [24]. Similarly, For the inframe deletion of metEF (Sp_0585-Sp_0586), a construct was created in which 921 bp of flanking DNA 59 to the Sp_0585 ATG (primers Sp0583F and Cm-Sp0584R) and 725 bp of flanking DNA 39 to the Sp_0586 ORF (primers Cm-Sp0587F and Sp0587R) were amplified by PCR from S. pneumoniae genomic DNA and fused with the chloramphenicol resistance marker (cat, amplified from pID701, a suicide vector containing cat gene, with primers CmF and CmR) by overlap extension PCR. The constructs were transformed into S. pneumoniae by homologous recombination and allelic replacement using competence stimulating peptide (CSP-1) (kind gift from D. Morrison) and selection with antibiotics according to established protocols [1,5]. Deletion of metQ and/or metEF was confirmed by PCR and sequencing of the PCR products (performed by Lark Technologies Inc. UK or UCL sequencing services using the Big Dye TM Terminator technique and gene specific PCR primers).

Cloning, expression and purification of His 6 -MetQ
Recombinant MetQ protein was expressed in E. coli and purified using an N terminal His-tagged QIAexpressionist TM system. Primers Sp0149.1 and Sp0149.2 amplified a full-length metQ from the S. pneumoniae D39 strain (excluding the 59 portion encoding the predicted N-terminal signal peptide), which was then ligated into the pQE30 expression vector to make the plasmid pPC138 and transformed into E. coli strain M15. For synthesis of recombinant proteins, the transformed cells were grown aerobically in 5 ml of Lennox broth (LB) for 8 h, and used to inoculate 50 ml LB for overnight aerobic growth at 376C. This was then used to inoculate 625 ml of LB at 306C. Cells were allowed to grow aerobically to an A 650 of 0.4-0.6 before production of recombinant protein was induced with 1 mM isopropyl 1-thio-b-Dgalactopyranoside (IPTG), followed by four hours of incubation at 306C. Cells were harvested by centrifugation at 44306 g for 15 min at 46C and resuspended in buffer A (50 mM potassium phosphate, pH 7.8, 500 mM NaCl, 10 mM immidazole and 20% glycerol). The cells were lysed by sonication and cell debris was removed by centrifugation at 380006 g for 30 min at 46C and the supernatant containing overexpressed protein was collected. Nickel affinity chromatography was used for the purification of the hexahistidine tagged protein using 5 ml His-trap column [GE Healthcare]. The soluble fraction containing the over produced protein was passed over column pre-equilibrated with buffer A. Weakly bound contaminants were removed by washing the column with 20 CV of buffer A containing 40 mM imidazole and the recombinant protein was eluted with elution buffer containing 500 mM imidazole. Purification products were analysed by SDS-PAGE gel and shown to contain 99% pure protein of the expected size for His 6 -MetQ. Purified protein was dialysed in 50 mM potassium phosphate pH 7.8 and concentrated using Vivaspin 5 kDa MWCO concentrators. Protein concentration was determined from the absorbance at 280 nm (A 280 ) using a calculated molar extinction coefficients of 45780 cm 21 M 21 for MetQ. His 6 -Sp0749 (LivJ) was also purified using an N terminal His-tagged QIAexpressionist TM system as described in [7].

Preparation of ligand free His 6 -MetQ
Binding proteins often co-purify with their cognate ligands and thus must be made ligand free in order to check binding with alternate ligands in vitro. Copurified ligand was removed from MetQ using guanidinium HCl (GdmHCl) as described previously [44]. The protein was partially unfolded while bound to the Histrap column, by including additional wash steps (in buffer A with 30 mm imidazole) of 40 CV with 2 m GdmHCl, 4 CV with 1.5 m GdmHCl, 4 CV with 1 m GdmHCl, 4 CV with 0.5 m GdmHCl, and finally 8 CV with 0 m GdmHCl. The protein was eluted with elution buffer as described previously.
Immunoblots and Triton-X-114 extract preparation S. pneumoniae were grown in THY medium until the OD 580 reached 0.6 and the cells were centrifuged, the pellet resuspended in sterile phosphate buffered saline (PBS) and sonicated. The cell lysate was centrifuged, and the supernatants used for further analysis by immunoblotting. Lipid associated proteins were extracted according to Khandavilli et al [24] using Triton X-114. Protein samples from Triton X-114 extracts were separated on SDS-PAGE 10% and 12% resolving gels respectively, blotted onto nitrocellulose membranes and probed with polyclonal mouse antisera raised against MetQ (1:2500 dilution) according to standard protocols [45].
Enzyme Linked Immunosorbent assay (ELISA) ELISA was performed according to Jomaa et al [31]. Microtitre ELISA plates were coated with 100 ml of antigen at a concentration of 5 mg ml 21 in TSA buffer and incubated overnight at 4uC. The plates were washed 5 times with ELISA wash buffer and soaked in 150 ml of 2% BSA-Tween for 2-4 hours at 37uC and washed again for 3-4 times with ELISA wash buffer. A 1/1000 dilution of mouse serum was made in BSA-Tween diluent buffer and added to the first well, and then 2 fold dilutions transferred to subsequent wells and incubated at 37uC for 4 hours. The wells were then washed 6 times with ELISA wash buffer and incubated overnight at 4uC in 100 ml of 1/15000 dilution of goat anti-mouse IgG conjugated to alkaline phosphate (Sigma) diluted in enzyme diluent. The ELISA plate was then washed 5 times in ELISA wash buffer and 100 ml of dinitrophenol (dNP) substrate solution added to each well and incubated for 1 hour at 37uC (dNP substrate solution was prepared by adding 1 tablet of dNP (Sigma) in 5 ml of water). The absorbance was read at 405 nm and antibody titre was calculated as the lowest dilution giving an OD equal to or greater than 0.3.

Opsonophagocytosis assays
To assess the effect of anti-metQ antibodies on the interaction of S. pneumoniae with phagocytes, we measured the proportion of a neutrophil cell line associated with fluorescent bacteria using a previously described flow cytometry opsonophagocytosis assay [31,46]. The complement source used was commercially available baby rabbit serum (Sigma S7764; rabbit HLA-ABC). S. pneumoniae strains were fluorescently labelled by incubation with 5, 6carboxyfluorescein-succinimidyl ester (FAM-SE; Molecular Probes, Eugene, Oreg.) solution (10 mg/ml in dimethyl sulfoxide; Sigma) in 0.1 M sodium bicarbonate buffer for 1 h at 37uC and then washed six times with Hanks balanced salt solution (HBSS) in 0.2% bovine serum albumin and stored in aliquots at 270uC in 10% glycerol (approximately 10 9 CFU/ml). The human cell line HL-60 (promyelocytic leukemia cells; CCL240; American Type Culture Collection, Manassas, Va.) was used to provide the effector cells after differentiation into granulocytes by using previously described protocols. Differentiated HL60 cells were harvested by centrifugation (1606 g, 8 min, 4uC) and washed twice with HBSS and once with HBSS in the presence of Ca 2+ and Mg 2+ . FAM-SE-labelled bacteria (10 6 CFU) were opsonized with 1/100, 1/40, and 1/10 dilutions of serum in a 96-well plate for 20 min at 37uC with horizontal shaking (170 rpm). Negative controls were included, using the same volume of HBSS. HL60 cells (10 5 ) were added to the opsonized bacteria in the microplate plate and incubated for 30 min at 37uC with shaking, after which the bacteria and cells were fixed using 3% paraformaldehyde and analyzed using flow cytometry. A minimum of 6,000 cells per sample were analyzed.

C-methionine uptake assays
Radioactive uptake assays were performed by the rapid filtration method as previously described [27] with minor modifications. S. pneumoniae strains were grown in C+Y medium until the OD 620 reached 0.2-0.4. These experiments were performed using a capsular serotype 2 S. pneumoniae strain (D39) as the mucoid colonies of the capsular serotype 3 strain prevented effective pelleting of the bacteria for these assays. Bacteria were harvested at 13000 g for 20 minutes and resuspended in 50 mM potassium phosphate buffer (pH 7.2) with 1 mM MgCl 2 to an OD 620 between 0.8-1.1. Uptake of methionine was determined in 1 ml assays containing 0.85 ml of bacterial cells and a final concentration of 50 or 750 mM methionine with 0.125 mCi 14 Cmethionine (GE Healthcare, United Kingdom). Samples (150 ml) containing bacteria, radioactive and non-radioactive substrates were removed at time intervals (0, 1, 2, and 3 minutes) and immediately filtered through glass fibre filters (Whatman GF/F), washed twice with 50 mM potassium phosphate buffer, and then washed filters placed in scintillation vials in 5 ml of Ready Safe scintillation cocktail (Beckman Coulter) and radioactivity determined using a Wallac 1214 RackBeta liquid scintillation counter. Bichinchoninic acid protein assay (Sigma Aldrich, UK) determined that an optical density (620 nm) of 1 was equivalent to 0.238 mg protein, and this figure was used to convert the radioactivity counts to nmol solute per mg protein.

Fluorescence Spectroscopy
Protein fluorescence experiment used a FluoroMax 4 fluorescence spectrometer (Horiba Jobin-Yvon) with connecting water bath at 376C. Ligand free protein (MetQ) was used at a concentration of 0.05 mM in 50 mM potassium phosphate, pH 7.8. Protein was excited at 281 nm with slit widths of 5 nm and emission was monitored at 352 nm with slit widths of 5 nm. To examine alternative ligands for MetQ, potential ligands were added at concentrations up to 5 mM. To determine the K D for ligand binding, the protein was titrated with increasing concentrations of the ligand and the corresponding fluorescence change was monitored in time acquisition mode. The cumulative fluorescence change was plotted in SigmaPlot and the K D was calculated from the hyperbolic fit of the binding curve. L-methionine, Dmethionine, DL-Homocysteine, a-Methyl-DL-methio-nine, L-Cysteine and Glycine were purchased from Sigma-Aldrich.

Isothermal Titration Calorimetry
Calorimetry experiment was performed using a VP-ITC instrument (MicroCal Inc., GE Heath Sciences) by taking 1.4 ml of ligand free MetQ in the calorimeter cell and 400 ml of L-methionine ligand in the syringe. The concentration of ligand in the syringe was typically 10 times that in the cell, whereas the cell concentration was chosen according to c value of 50, where c = [macromolecule]/(predicted) K D . Experiments were carried out in 50 mM potassium phosphate buffer, pH 7.8, at 376C. Both cell and syringe solutions were degassed at 356C for 10 min before use. The titrations were performed as follows. A single preliminary injection of 3 ml of ligand solution was followed by 40 injections (6 ml), delivered at an injection speed of 10 ml s 21 . Injections were made at 3-min intervals with a stirring speed of 307 rpm. Raw titration data were integrated and fit to a one-site model of binding using MicroCal Origin version 7.0.

Animal models of infection models
Infection experiments were performed in age and sex matched groups of outbred CD1 mice (Charles River Breeders) between 4 to 8 weeks old. For mixed infections, equivalent numbers of bacteria from stocks of wild-type and mutant S. pneumoniae strains were mixed and diluted to the appropriate concentration. For the nasopharyneal colonisation model, 10 7 CFU of bacteria in 10 ml were administered by intranasal inoculation under halothane general anaesthesia, and nasal washes were obtained 5 days post colonisation. Serial dilutions of the samples were plated onto Columbia blood agar plates containing optochin (50 mg ml 21 ) and/or gentamycin (5 mg ml 21 ) to differentiate pneumococcus from other contaminating streptococci and to enumerate CFU. The CI was calculated using the formula: ratio of mutant to wild-type strain recovered from mice divided by the ratio of mutant to wild-type strain in the inoculum [47]. A CI of less than 1 indicates the mutant strain is attenuated in virulence compared to the wildtype strain, and the lower the CI the more attenuated the mutant strain. Experiments were performed using pure inocula of wildtype or mutant strains to calculate bacterial CFU in recovered target organs for each strain at specific time points [47]; for the systemic model of infection 1610 3 CFU bacteria in 100 ml were inoculated by intraperitoneal (IP) injection and spleen homogenates obtained at 24 hours for plating [5,6,24]; and for the pneumonia model, 5610 6 CFU bacteria in 40 ml were given by intranasal (IN) inoculation under halothane general anaesthesia, and lung homogenates obtained at 48 hours for plating [6,24]. For the vaccination infections, mice were given by IP injection 100 mg of MetQ or just PBS mixed with 10% alum on three occasions separated by 7 to 8 days. Serum was obtained for immunoblots and antibody titres ELISAs from five mice four weeks after the last vaccination, and the remaining vaccinated mice challenged by IP inoculation with 1610 4 S. pneumoniae D39 strain CFU. Disease progression was monitored using previously established criteria to identify mice likely to progress to fatal disease [5].

Statistical analysis
All the in vitro growth curves were performed in triplicates and represented as means and standard deviations. Results of growth curves, radioactive uptake and binding assays were analysed using two-tailed t-tests. Target organ CFU were compared between strains using the Mann Whitney U test, and the survival of infected mice using the log rank test.