Present address: Université Paul Sabatier-CNRS, LMGM, Bât. IBCG, 118 route de Narbonne, F-31062 Toulouse Cedex 9, France.
The global nutritional regulator CodY is an essential protein in the human pathogen Streptococcus pneumoniae
Article first published online: 29 AUG 2010
© 2010 Blackwell Publishing Ltd
Volume 78, Issue 2, pages 344–360, October 2010
How to Cite
Caymaris, S., Bootsma, H. J., Martin, B., Hermans, P. W. M., Prudhomme, M. and Claverys, J.-P. (2010), The global nutritional regulator CodY is an essential protein in the human pathogen Streptococcus pneumoniae. Molecular Microbiology, 78: 344–360. doi: 10.1111/j.1365-2958.2010.07339.x
- Issue published online: 12 OCT 2010
- Article first published online: 29 AUG 2010
- Accepted manuscript online: 12 AUG 2010 12:00AM EST
- Accepted 2 August, 2010.
- Top of page
- Experimental procedures
- Supporting Information
CodY is a global regulator highly conserved in low-G+C Gram-positive bacteria. It plays a key role in the adaptation of Bacillus subtilis to nutritional limitation through repression of a large gene set during exponential growth and relief of repression upon starvation. In several pathogenic bacteria, CodY regulates major virulence genes. Our interest in Streptococcus pneumoniae CodY originates from our observations that the oligopeptide permease Ami was involved in repression of competence for genetic transformation. We hypothesized that peptide uptake through Ami feeds amino acid pools, which are sensed by CodY to repress competence. As our initial attempts at inactivating codY failed, we launched an in-depth analysis into the question of the essentiality of codY. We report that codY cannot be inactivated unless a complementing ectopic copy is present. We obtained genetic evidence that a recently published D39 codY knock-out contains additional mutations allowing survival of codY mutant cells. Whole-genome sequencing revealed mutations in fatC, which encodes a ferric iron permease, and amiC. This combination of mutations was confirmed to allow tolerance of codY inactivation. The amiC mutation is in itself sufficient to account for the strong derepression of competence development observed in D39 codY cells.
- Top of page
- Experimental procedures
- Supporting Information
CodY is a global regulator highly conserved in low-G+C Gram-positive bacteria (Sonenshein, 2005). In Bacillus subtilis, the CodY regulon is large, encompassing nearly 200 genes (Molle et al., 2003), most of which are repressed during exponential growth and induced when cells experience nutrient deprivation. In several pathogenic bacteria, CodY regulates major virulence genes (for review, see Sonenshein, 2005; 2007). The repressor function of CodY (i.e. its DNA-binding activity) is activated by interaction with branched-chain amino acids (BCAAs), as originally shown in vivo in Lactococcus lactis (Guédon et al., 2001). In species other than the streptococci, lactococci and enterococci, CodY is also activated by interaction with GTP as first demonstrated (Ratnayake-Lecamwasam et al., 2001) and further documented (Handke et al., 2008) with the B. subtilis protein. In the latter bacterium, CodY is known to control not only metabolic pathways but also cellular processes, such as motility, sporulation and competence for genetic transformation (Sonenshein, 2005; 2007). In the latter case, CodY represses both comK, which encodes the master transcriptional activator of competence (com) genes, and srfA, a key operon for transcriptional activation of comK (Serror and Sonenshein, 1996).
Until recently, little was known regarding the role(s) of CodY in Streptococcus pneumoniae. Our interest in CodY stems from our previous observations that the oligopeptide permease Ami was, together with its dedicated oligopeptide-binding lipoproteins (Obl, i.e. AmiA, AliA and AliB) (Alloing et al., 1994), indirectly involved in the regulation of competence (Claverys et al., 2000; Claverys and Håvarstein, 2002). Competence development of an obl mutant thus occurred at a ∼50-fold reduced cell density compared with wild type (Alloing et al., 1998). We hypothesized that the uptake of oligopeptides plays a key role in metabolic regulation in S. pneumoniae, by providing information on nutrient availability (Claverys et al., 2000). Exhaustion of nutrients would be sensed through a mechanism involving peptide uptake by the Ami-Obl oligopeptide permease, followed by peptidase digestion to release amino acids (aa) leading to replenishment of aa pools, which in turn affect a global regulatory protein such as CodY (Claverys et al., 2000). The above-mentioned demonstration that BCAAs are effectors of CodY in several species would fit nicely with this hypothesis since S. pneumoniae Obl proteins have different but overlapping specificities for peptides including those containing BCAAs (Alloing et al., 1994). Thus import of BCAAs (as part of oligopeptides) by the Ami-Obl permease could directly impact CodY activity. This situation would not be unprecedented since CodY-dependent repression of comK and srfA was lost in an opp (oligopeptide permease) mutant of B. subtilis (Serror and Sonenshein, 1996). To account for the negative impact of Ami-Obl on competence induction, we explicitly stated in our model that CodY should repress competence (Claverys et al., 2000).
To test this model, we initiated experiments aimed at inactivating S. pneumoniae codY. Surprisingly, this gene turned out to be difficult to inactivate. However, codY inactivation had recently been achieved through insertion of a trimethoprim resistance gene (trim) in strain D39 (Hendriksen et al., 2008). Microarray analysis using this mutant suggested that pneumococcal CodY functions mainly as a transcriptional repressor, as 43 of the 47 genes differentially expressed in the codY::trim mutant were found to be upregulated. The inefficiency in constructing a codY knock-out in strain R6 prompted us to launch an in-depth analysis of the question of the essentiality of codY. Here, we present our data establishing that codY is essential in our laboratory strains, as well as in S. pneumoniae R6 and encapsulated D39 strains. We provide genetic evidence that, in addition to the codY knock-out, the codY::trim mutant used for transcriptome analysis of the CodY regulon and for virulence studies (Hendriksen et al., 2008) contains additional mutations (suppressors) allowing survival of codY mutant cells. We document a strong derepression of competence in such codY mutant cells. Finally, we use comparative whole-genome sequencing to identify suppressor mutations in fatC, which encodes a ferric iron permease, and amiC. We confirm that this combination of mutations allows tolerance of codY inactivation and discuss possible reasons for CodY essentiality in S. pneumoniae in light of these observations.
- Top of page
- Experimental procedures
- Supporting Information
Minitransposon insertion mutagenesis of S. pneumoniae codY is not possible
A codY PCR fragment amplified with the primer pair MP188–MP189 (Table 1) was used as target for mariner mutagenesis (Experimental procedures). Analysis of more than 50 minitransposon insertions through PCR revealed that only six clones produced an MP188–MP189 fragment of which the size was consistent with that predicted for integration of the kanR minitransposon (donated by plasmid pR410; Table 1). Localization of these six insertions indicated that none of them had inserted into the codY gene (Fig. 1A). Given our previous experience with mariner mutagenesis of the ciaRH (Martin et al., 2000), endA, comEAC, comFAC, comGAB and dprA (Bergéet al., 2002), as well as cibABC (Guiral et al., 2005), comM–lytR (Håvarstein et al., 2006) and radC (Attaiech et al., 2008) loci, we concluded that the failure to isolate insertions in a locus occupying a central position on the targeted PCR fragment and the biased distribution observed were strongly indicative of the essentiality of codY, at least in the genetic background and under the plating conditions (CAT agar) used. [It is of note that a similar conclusion regarding the essentiality of lytR (spr1759) based on the failure to isolate mariner insertions in this locus (Håvarstein et al., 2006) was confirmed in a further study (Johnsborg and Håvarstein, 2009).]
|D39||Serotype 2||NCTC 7466|
|D39ΔcodY||D39 ΔcodY::trim (socY*)b; TrimR, (MtxR)b||Hendriksen et al. (2008)|
|D39Δcps||D39 Δcps::kan; KanR||Hendriksen et al. (2008)|
|D39ΔcpsΔcodY||D39 Δcps::kanΔcodY::trim (socY)c; KanR, TrimR, (MtxR)c||Hendriksen et al. (2008)|
|R6||Unencapsulated derivative of D39||Burghout et al. (2007)|
|R246||R800 but hexAΔ3::ermAM; EryR||Mortier-Barrière et al. (1998)|
|R304||R800 derivative, nov1, rif23, str41; NovR, RifR, SmR||Mortier-Barrière et al. (1998)|
|R800||R6 derivative||Lefèvre et al. (1979)|
|R895||R800 but ssbB::luc (ssbB+); CmR||Chastanet et al. (2001)|
|R1501||R800 but comC0||Dagkessamanskaia et al. (2004)|
|R1818||R1501 but hexAΔ3::ermAM; EryR||This study|
|R2349||R1501 but CEPM–codY (the resulting duplication of codY is denoted codY+/+); KanR||This study|
|R2350||R895 but codY+/+; KanR, CmR||This study|
|R2424||R2350 but genuine codY inactivated by mariner insertion spc3A (the resulting codY combination is denoted codYspc3/+); KanR, CmR, SpcR||This study|
|R2425||R2350 but ectopic codY inactivated by mariner insertion spc3A (the resulting codY combination is denoted codY+/spc3); KanR, CmR, SpcR||This study|
|R2427||R6 but codY+/+; KanR||This study|
|R2428||R1818 but codY+/+; EryR, KanR||This study|
|R2430||R895 but pmalR (pAPM22); CmR, EryR||This study|
|R2432||R2424 but pmalR (pAPM22) ; CmR, EryR, KanR, SpcR||This study|
|R2437||R2349 but str41; KanR, StrR||This study|
|R2438||R2437 but codYspc3/+; KanR, SpcR, StrR||This study|
|R2549||R895 but amiC9; CmR, MtxR||This study|
|R2641||R895 but kan90C::codY+,str41; CmR, SpcR, StrR||This study|
|R2644||R2641 but spc90C::codY+; CmR, SpcR, StrR||This study|
|R2737||R895 but CEPM; CmR, KanR||This study|
|R3002||R246 but fatC::cat23C (from TD131); CmR, EryR||This study|
|R3003||R246 but amiC9; EryR, MtxR||This study|
|R3004||R3002 but amiC9; CmR, EryR, MtxR||This study|
|TD73||D39 but codY+/+; KanR||This study|
|TD80||D39ΔcpsΔcodY (socY)b but str41; KanR, StrR, TrimR||This study|
|TD81||TD80 but codY::spc3A; KanR, SpcR, StrR||This study|
|TD82||D39Δcps but ssbB::luc (ssbB+); CmR, KanR||This study|
|TD83||D39ΔcpsΔcodY but ssbB::luc (ssbB+); CmR, KanR, TrimR||This study|
|TD84||TD80 but amiC9; CmR, KanR, MtxR||This study|
|TD95||TD83 but spc90C::codY+; CmR, KanR, SpcR||This study|
|TD96||TD95 but codY+; CmR, KanR||This study|
|TD129||D39Δcps but fatC::spc23C; KanR, SpcR||This study|
|TD130||TD129 but amiC9; KanR, MtxR, SpcR||This study|
|TD131||TD129 but fatC::cat23C; CmR, KanR||This study|
|TD135||TD130 but ssbB::luc (ssbB+); CmR, KanR, MtxR, SpcR||This study|
|TD138||TD129 but ssbB::luc (ssbB+); CmR, KanR, SpcR||This study|
|TD154||TD135 but ΔcodY::trim; CmR, KanR, MtxR, SpcR, TrimR||This study|
|pAPM22||pLS1 derivative carrying the malR gene; EryR||Puyet et al. (1993)|
|pCEP||pSC101 derivative (i.e. low-copy-number plasmid) carrying CEP; SpcR, KanR||Guiral et al. (2006)|
|pCEP2||pKL147 derivative (i.e. high-copy-number plasmid) carrying an EcoRI/PstI fragment from pCEP; SpcR, KanR, ApR||This study|
|pCEP2–codY||pCEP2 derivative carrying codY under PM control; SpcR, KanR, ApR||This study|
|pEMcat||ColE1 derivative carrying a CmRmariner minitransposon; ApR, CmR||Akerley et al. (1998)|
|pKL147||pUS19 derivative containing gfpmut2 fused to the 3′ end of dnaX with a linker; SpcR, ApR||Lemon and Grossman (1998)|
|pR410||pEMcat derivative carrying a KanR (kan gene) mariner minitransposon; ApR, KanR||Sung et al. (2001)|
|pR412||pEMcat derivative carrying an SpcR (aad9 gene, also called spc) mariner minitransposon; ApR, SpcR||Martin et al. (2000)|
|ami1||GCGCAAACAGGCTCTAAGGG; amiA; +1815||This study|
|ami2||TCAGGAATTCCTGCTGCCATTAT; amiC; +1257||This study|
|ami4||CCTGACTCACCTACCAAGGCTA; amiD; +712||This study|
|ami5||CCTTCACCGAAGGAAATTTCTA; amiE; +121||This study|
|ami6||TTAGCTGACTTCAACCCACTACA; amiF; +1027||This study|
|amiF1||GCCTTGCTTTCAGCGGTACCAAT; amiF; +789||This study|
|AM40||AGAGTTTCGGATGGTTTGGA; treR; +347||This study|
|codY1||CAAGGATCAGTTTTCCCATATTTTCG; codY; +1636||This study|
|codY2||CTTCGTGTCCTTCGTGACTTTA; codY; −1004||This study|
|codYatg||tgaatcATGaCACATTTATTAGAAAAAACTAG; codY; 0||This study|
|codYstop||aaattggatccTTTGTCATTAGTAATCTCTTTTC; codY; +797||This study|
|fat1||GCGAACGAATGATTTACTGG; fatD; −659||This study|
|fat2||TCTCACCAGTCTTTCCACCC; fatB; +1388||This study|
|fatC1||TAAAAGCAAACATACCAAGC; fatC; −9||This study|
|fatC2||TAAAGAATAAGAAGCCACCC; fatC; +909||This study|
|HBDamiCF1||ACGGCTGATAAACGTGATAA; amiC; +145||This study|
|HBDamiCF2||GTCGTTGGTCTTGTCTTCAT; amiC; +1381||This study|
|HBDamiCR1||TAAATTCTCCCAAAGTCCAA; amiC; +343||This study|
|HBDamiCR2||CGCATCAATAGTTTCAGAGG; amiC; +1568||This study|
|HBDfatCF||ACACTGATGAAGCAAGACCT; fatC; +376||This study|
|HBDfatCR||CAATATCTGAGCCGTTTCTC; fatC; +645||This study|
|kan1||ATCATGTCCTTTTCCCGTTCCAC; kan; +191||This study|
|MP127||CCGGGGACTTATCAGCCAACC; mariner transposon||Martin et al. (2000)|
|MP128||TACTAGCGACGCCATCTATGTG; mariner transposon||Martin et al. (2000)|
|MP188||TTCATTTTCACCAACCAGGTTAC; codY; +1032||This study|
|MP189||ATTGGCTGCTGAGTTTACTCCAG; codY; −618||This study|
|MP192||ggatccACGTCATCAACTAAATAGCG; aliA; −343||This study|
|MP193||CAGAAGCTTTCTGGTTTGTT; aliA; −539||This study|
|MP194||TTGGAATTCCCTCTTCTGGAAC; dexB; +908||This study|
|MP195||ttagttgatgacgtggatccGCTTTTTATACAGTCCTCCC; dexB; +1693||This study|
|rpsL_3||TGACATGGATACGGAAGTAG; rspL; −798||This study|
|rpsL_4||ATGGTAAGCTGAGTTATAGC; rpsL; +1204||This study|
Construction of a strain harbouring a second (ectopic) copy of codY
The failure to inactivate codY through mariner mutagenesis prompted us to construct a strain harbouring a second copy of codY that should tolerate inactivation of one of the two copies. The second copy was inserted at CEP (chromosomal expression platform) (Guiral et al., 2006) under the control of the maltose-inducible promoter, PM (Fig. 1B). The resulting strain is referred to as codY+/+ hereafter. The duplication of codY had no detectable effect on growth and spontaneous competence induction when cells were grown with 1% maltose, i.e. under conditions leading to full induction of the PM promoter (Guiral et al., 2006), suggesting that the level of CodY attained with this induction system is not detrimental to the cell (data not shown).
Minitransposon insertions in codY are readily obtained in diploid (codY+/+) cells
A codY+/+ strain (R2349) was then used as recipient for mariner mutagenesis of codY with plasmid pR412 as donor of minitransposon (spcR cassette; Table 1). In contrast to the failure to inactivate codY using wild-type recipient cells, minitransposon insertions were readily obtained in cells grown in maltose. A codY1–codY2 (Table 1) PCR fragment of which the size was consistent with that predicted for integration of the spcR minitransposon was observed for 11 out of 20 randomly selected clones. Five of these insertions turned out to inactivate codY as judged from their location (Fig. 1C) and from the inability to introduce them at high frequency in a wild-type recipient (data not shown but see next section).
The isolation of minitransposon insertions in the genuine copy of codY suggested that the ectopic copy placed at CEP was able to complement codY deficiency. This conclusion was further strengthened by the observation that upon transformation of codY+/+ cells with a PCR fragment carrying the codY::spc3A insertion (generated with the primer pair codYatg–codYstop; Fig. 1C and Table 1), SpcR transformants distributed about equally between the genuine codY (e.g. strain R2424 the genotype of which is denoted codYspc3/+; Fig. 2A) and the ectopic CEPM–codY loci (e.g. strain R2425 the genotype of which is denoted codY+/spc3; Fig. 2B).
Transformation frequencies of codYspc3/+ in wild-type and codY+/+ cells confirm codY essentiality
To unambiguously demonstrate the essentiality of codY in our strains, we then used chromosomal DNA from a codYspc3/+ strain (R2438) as donor in transformation of a wild-type recipient. If codY is essential, the survival of transformants harbouring the codY::spc3A knock-out mutation is predicted to rely on the simultaneous integration of CEPM–codY–kan (i.e. a double transformation event occurring independently as the two loci are not genetically linked). On the other hand, a codY+/+ recipient should readily accept the codY knock-out. Transformation frequency should therefore be significantly reduced in wild-type cells compared with a codY+/+ recipient. In full agreement with this prediction, transformation of codY::spc3A occurred in codY+/+ (strain R2350) cells with a frequency close to that of the str41 (StrR) reference marker (note that transformation of the spcR cassette, which requires integration of ∼1.1 kb heterologous DNA, is expected to occur with a two- to threefold reduced frequency compared with the point-mutation reference marker str41) (Fig. 3A). In contrast, ∼20-fold reduction in the frequency of SpcR transformants was observed with wild-type (R895) cells (Fig. 3A). The observed frequency was close to that calculated as the product of individual transformation frequencies of SpcR and KanR. To verify that SpcR transformants obtained with the wild-type recipient had simultaneously acquired resistance to Kan, 10 SpcR transformants were isolated. All of them turned out to carry the CEPM–codY–kan construct (data not shown). Altogether, these data demonstrated that codY is an essential gene in S. pneumoniae R800 and its derivatives.
codY is also essential in R6 and in encapsulated D39
R800 was originally derived from R6 through introduction of a suppressor mutation (originating from strain Cl3, another D39 derivative) (Tiraby et al., 1975) that greatly improved growth of ami mutants (Lefèvre et al., 1979). This led us to wonder whether the presence of this uncharacterized suppressor was responsible for the essentiality of codY. We therefore introduced the CEPM–codY construct in R6 (generating strain R2427) and measured transformation frequencies of codY::spc3A using codYspc3/+ chromosomal DNA as donor. Yield of SpcR transformants was significantly lower in R6 than in its codY+/+ derivative indicating that codY is essential in this background as well (Fig. 3B).
Inactivation of codY was previously reported in the encapsulated S. pneumoniae D39 strain (Hendriksen et al., 2008). To check whether our negative data could be explained by the use of unencapsulated strains, of different medium and/or plating conditions, we first tried to transform D39 with an amiF1–kan1 PCR fragment carrying codY::spc3A (Fig. 2B), using previously described conditions for plating of codY- transformants on Columbia base agar (Hendriksen et al., 2008). No SpcR transformants (i.e. < 10−6) could be obtained (data not shown). We therefore adopted the same strategy as the one described above for R6. Transformation of codY::spc3A using codYspc3/+ chromosomal DNA as donor yielded ∼20-fold fewer SpcR transformants in D39 than in its codY+/+ derivative (strain TD73) indicating that introduction of a codY knock-out in D39 also requires the simultaneous transfer of the complementing ectopic CEPM–codY gene (Fig. 3C). In full agreement with this interpretation, SpcR transformants were also KanR. We concluded that codY is also essential in encapsulated D39 and that this gene must therefore be added to the list of essential pneumococcal genes (Thanassi et al., 2002).
Re-investigation of the previously described codY knock-out reveals the presence of additional suppressor mutations or a chromosomal rearrangement
We then re-examined the previously constructed codY::trim mutant, strain D39ΔcodY (Hendriksen et al., 2008), as the transformation frequency upon mutant generation was very low (W. Hendriksen, pers. comm.). PCR analysis of the codY chromosomal region of strain D39ΔcpsΔcodY, which was derived from D39ΔcodY by transformation with a PCR fragment harbouring the Δcps::kan cassette, as previously described (Bootsma et al., 2007), confirmed the published structure (data not shown). Preliminary attempts using D39ΔcpsΔcodY chromosomal DNA as donor revealed that transformation of the codY::trim construct into R800 derivatives occurred only at very low frequency (data not shown). To facilitate the comparison with transformation experiments reported above, we replaced the codY::trim construct with the codY::spc3A insertion (using as donor a codY1–codY2 PCR fragment) and we introduced the str41 reference marker (using as donor a PCR fragment amplified with the rpsL_3–rpsL_4 primer pair; Table 1) thus generating strain TD81. Then, using TD81 chromosomal DNA as donor, we compared transformation frequencies of codY::spc3A in the same pairs of codY+ and codY+/+ derivatives of R800, R6 and D39 used in Fig. 3. Transformation frequencies in codY+ strains were reduced by 544-, 762- and 351-fold, respectively, in R800, R6 and D39, compared with their codY+/+ isogenic derivatives (Fig. 4A–C). These very large reductions in transformation frequency (compared with the ∼20-fold reduction observed when the simultaneous transfer of codY::spc3A and CEPM–codY from a codYspc3/+ donor was required; Fig. 3) indicated that transfer of the codY::spc3A from strain TD81 must be accompanied by the transfer of more than one point mutation. Alternatively, the very low transformation frequency could be accounted for by the requirement for a simultaneous chromosomal rearrangement. We concluded that the original D39ΔcpsΔcodY strain either contained two additional suppressor mutations that are presumably acting together to compensate for the absence of CodY and restore cell viability, or harboured a chromosomal rearrangement allowing survival in the absence of CodY. We tentatively named it/them socY (for suppressor of c odY).
To get a possible insight into the molecular nature of the mutation(s) involved, we carried out the same transformation experiments but with a pair of hex mutant derivatives of R800 (strain R1818 and its codY+/+ derivative, R2428). The Hex system of S. pneumoniae is known to correct out some mismatches at the donor–recipient heteroduplex stage in transformation (Claverys and Lacks, 1986). It is particularly efficient at correcting transition mismatches (i.e. A/C or G/T) as well as short frameshifts (Gasc et al., 1989). Transformation frequencies with hex- recipient cells were improved ∼4.5-fold (125- versus 544-fold reduction in the codY+ parent compared with the codY+/+ derivative, in hex- and hex+ strains respectively; Fig. 4D), which suggests that one of the socY mutations is recognized by the Hex system and is therefore possibly a transition or a short frameshift. Alternatively, the recombination event leading to the putative chromosomal rearrangement may involve the formation of a heteroduplex intermediate harbouring mismatches susceptible to the Hex system.
Effect of codY inactivation on spontaneous competence development
To characterize the role of S. pneumoniae CodY with respect to growth and the regulation of competence, we first tried to deplete CodY making use of a codYspc/+ strain, i.e. a strain in which the only functional copy of codY was under the control of the PM promoter. In the light of the failure to obtain a biologically significant depletion of CodY (supplementary results in Supporting information and Fig. S1), we chose to characterize a codY socY strain with respect to spontaneous competence development and growth. Competence was monitored throughout incubation at 37°C by using a transcriptional fusion of the luc gene, which encodes luciferase, to the ssbB gene. The latter is known to be specifically induced at competence. The ssbB::luc fusion thus reports on competence through light emission by luciferase (Prudhomme and Claverys, 2007). We introduced the ssbB::luc transcriptional fusion (using R895 chromosomal DNA as donor) in strains D39Δcps and D39ΔcpsΔcodY socY thus generating strains TD82 and TD83 respectively (Table 1). We then compared competence profiles of strains TD82 and TD83 during growth in C+Y medium with initial pH values between 6.48 and 7.26, since spontaneous competence induction is known to be strongly dependent on the initial pH. For instance, initial pH values between 6.8 and 8.0 affected the timing of occurrence and the level of competence (Chen and Morrison, 1987). While the wild-type parent developed spontaneous competence only in cultures with initial pH values above 7.0 (Fig. 5A and B), the codY mutant could develop competence under acidic conditions, down to an initial pH value of 6.70 (Fig. 5C and D). It is of note that despite the presence of the socY suppressor, codY mutant cells grew more slowly than wild-type cells in C+Y medium (Fig. 5A–D). codY mutant cells thus entered the stationary phase of growth after ∼270 min incubation compared with ∼170 min for wild-type cells. This > 50% increase in generation time might indicate that CodY plays an important role in the regulation of pneumococcal growth. Despite the fact that codY mutant cells grew more slowly than wild type, they developed competence at about the same time (e.g. maximum competence after 87 versus 82 min incubation at pH 7.26; Fig. 5A and C), which corresponded to OD492 values of 0.083 and 0.115 respectively. This observation was also consistent with upregulation of competence in the codY mutant. The strong competence-upregulated (cup) phenotype (Martin et al., 2000) displayed by codY mutant cells would be consistent with the hypothesis that CodY normally represses competence under acidic conditions in wild-type cells. However, the uncharacterized suppressor mutations (socY) in strain TD83 could also be responsible for the observed cup phenotype. This prompted us to investigate the phenotype of socY (codY+) cells.
Impact of socY on spontaneous competence development
First, to establish whether restoration of codY+ would be tolerated in a socY genetic background, strain TD83 was transformed with R2644 chromosomal DNA. R2644 carries the spc90C insertion immediately upstream of the codY+ gene (see Experimental procedures). Integration of the spcR cassette by transformation could thus be accompanied by the removal of the adjacent codY::trim cassette. A failure to survive of TrimS (i.e. codY+) excisants should result in a drastic reduction in the number of SpcR transformants, since recombination events leading to spc integration without the simultaneous deletion of trim are rare due to the reduced distance between the two cassettes (105 bp). The high SpcR to StrR transformant ratio observed (average value 0.32 ± 0.06) was consistent with the simultaneous occurrence of both integration (of the spc cassette) and excision (of the trim cassette) events, and suggested that socY (codY+) cells are viable. To establish this, four of four randomly chosen SpcR transformants were first checked to be TrimS. Then, they were shown to readily re-accept codY::trim when transformed with TD80 chromosomal DNA (TrimR/StrR average ratio of 0.25 ± 0.08 over 15 independent cultures). These data confirmed that the SpcR TrimS transformants were still socY and indicated that none of the two suppressor mutations or the putative chromosomal rearrangement is detrimental to pneumococcal cells when CodY is present. One of the SpcR TrimS clones was retained and named strain TD95. The spc90C cassette was then removed from strain TD95 by transformation with a PCR fragment generated on R800 chromosomal DNA with the MP188–MP189 primer pair, followed by phenotypic expression and segregation in liquid culture (C+Y medium) for 4.5 h, and plating on CAT agar without antibiotic. Individual colony screening was carried out to isolate an SpcS clone, TD96. Upon transformation of TD96 with TD80 chromosomal DNA, a TrimR/StrR ratio of 0.42 ± 0.07 was observed indicating that TD96 had remained socY.
The competence profile of TD96 during growth in C+Y medium with initial pH values between 6.48 and 7.26 was compared with that of TD82 (wild type) and TD83 (codY::trim socY). TD96 socY (codY+) cells grew more rapidly than parental codY::trim socY cells, but still more slowly than wild type (entry into stationary phase after ∼210 min versus ∼170 min; Fig. 5E and F). Interestingly, socY cells displayed a stronger cup phenotype than codY socY cells as judged first from their ability to develop spontaneous competence at pH 6.60 (Fig. 5F) and second, from the very early development of competence (36 min at pH 7.26) compared with both wild-type and codY socY cells (Fig. 5). Thus, the socY mutations alone confer a strong cup phenotype. The net effect of codY inactivation in this background is to attenuate this cup phenotype. From these data, it is therefore difficult to conclude that CodY acts as a repressor of competence in pneumococcal cells.
Whole-genome sequence comparison suggests fatC and amiC mutations suppress inviability of codY mutants
In an attempt to identify the socY (suppressor) mutations, whole-genome resequencing of strains D39 and D39ΔcodY was performed (Experimental procedures). Compared with the previously published D39 NCTC 7466 genome (Lanie et al., 2007), the D39ΔcodY strain had 14 mutations also found in its parent. As our D39 was obtained originally from NCTC (via T. Mitchell), these mutations presumably have arisen during laboratory cultivation. None of them appeared to alter an important function (Table S1), which is consistent with the fact that this D39 strain displayed full virulence in a mouse model (Hendriksen et al., 2008).
The codY mutant had also mutations flanking the trim cassette [a CT transition, a GTAGC frameshift and a TA transversion, respectively, at positions −175, +681/682 and +724 (positions are given with respect to the ATG of codY)]. The two point mutations occurred in between the trim cassette and the oligonucleotide primers used to amplify the codY region (Hendriksen et al., 2008), and were presumably introduced during polymerase chain reaction. The transition mutation is in itself sufficient to lower the frequency of integration of the codY::trim cassette by approximately fivefold during transformation of mismatch-repair-proficient strains.
Most relevant with respect to the question of CodY essentiality, the codY mutant had, in addition, a mutation in fatC and a second, variable mutation in amiC (Fig. 6A). The former mutation (a CT transition), fatCC496T changed a CAA (Gln) codon into TAA (stop) in a gene belonging to the fatD–fatC–fecE–fatB operon (spd_1649–1652 in D39; spr1684–1687 in R6; sp1869–1872 in TIGR4) (Fig. 6A, top). This operon [also called piuBCDA or pit1 (Brown et al., 2002)] encodes the major ferric iron/haem [fatB has been shown to bind haemin (Tai et al., 2003)] transporter of S. pneumoniae (Ulijasz et al., 2004). The fatCC496T change is predicted to result in the synthesis of a truncated FatC protein (165 instead of 318 aa) and was detected in 100% of forward and reverse reads. As concerns the base variations in amiC, which encodes a 598 aa protein and belongs to the amiACDEF operon (spd_1671-1667 in D39; spr1707-1703 in R6; sp1891-1887 in TIGR4), three types of mutations were observed: two single-base changes (#2 and #3 in Fig. 6A, bottom) and a more complex mutation (#1 in Fig. 6A, bottom). These variations were detected in only a fraction of sequence runs. Mutations #2 (amiCG1438T) and #3 (amiCG1459T) change a GAC (Asp) codon into TAC (Tyr) (aa 480 and 487 respectively); these mutations were detected in 30.8% of forward and 11.1% of reverse reads (total with variation: 19%), and in 57.1% of forward and 40.9% of reverse reads (total with variation: 47%) respectively. Interestingly, both changes affect an Asp aa conserved in the entire family of Opp (oligopeptide), Dpp (dipeptide) and App (nickel) permeases (our observations), which suggests that they are functionally important and that the corresponding AmiC proteins may have lost their activity. The complex mutation #1 consisted of a GTT frameshift (amiCG246TT) and the almost adjacent AC change (amiCA248C); both mutations shared the same sequencing parameters (detection in 25.0% of forward and 33.3% of reverse reads), which strongly suggests that they resulted from a single mutational event (total with variation, 29%). This mutation could thus be described as a GGATTGC change. [In fact, this change had already been observed during analysis of mismatch repair specificity in S. pneumoniae and referred to as amiA29 (at this time, the ami locus was thought to be a single gene, hence the name amiA) (Gasc et al., 1989) (for a molecular explanation of this mutational event, see Fig. S2).] This complex mutation is predicted to result in the synthesis of a truncated protein harbouring the first 81 aa of AmiC fused to 13 ‘new’ aa (resulting from the frameshift).
At first sight, the presence of changes #1, #2 and #3 in only a fraction of sequence runs appeared puzzling and suggested their late occurrence in an original ΔcodY fatCC496T clone. As a first attempt to establish whether an ami mutation was required for tolerance of codY, we analysed the ami locus of D39ΔcpsΔcodY, considering it a subclone of the original D39ΔcodY strain since it was generated by transformation of this strain with a PCR fragment harbouring the Δcps::kan cassette. Taking advantage of the fact that amiC–F mutations normally confer resistance to methotrexate (MtxR), transformation of an MtxS strain (R1501) with a series of PCR fragments covering the ami region of strain D39ΔcpsΔcodY was used to localize any ami mutation (Fig. S3). Transformation data were consistent with the presence of mutation #2 or #3 in strain D39ΔcpsΔcodY; DNA sequencing confirmed that this strain had inherited the latter mutation (amiCG1459T). These data prompted us to directly confirm the mixed structure population with respect to amiC of the original D39ΔcodY strain. An aliquot of the original stock was plated, 16 individual clones were picked and amiC (and fatC) regions were sequenced (Table S2). All clones turned out to harbour one of the three amiC mutations identified during whole-genome sequencing (and 11 out of 11 clones sequenced carried the fatC mutation). Interestingly, the overall distribution between the three types was in very good agreement with that inferred from the analysis of genome sequence data (Table S2). It is also of note that three clones harboured change #1 (i.e. both amiCG246TT and amiCA248), which provided support to the view that a single molecular event accounts for a complex change (Fig. S2). Altogether, these results strongly suggested that inactivation of amiC was required for tolerance of codY inactivation. We tentatively concluded from these observations that the combination of fatC and amiC mutations suppresses inviability of codY mutants.
The fatC amiC combination allows tolerance of codY inactivation
To confirm this conclusion, we generated mariner insertion mutants in fatC (Fig. S4). Then, taking advantage of the availability of the well-characterized amiC9 mutation (amiCC355T) [previously named amiA9 (Gasc et al., 1989)] which changes a CAA codon (Gln) into a TAA (stop) codon leading to the synthesis of a truncated AmiC protein (118 aa), we investigated the ability of recipient cells harbouring the fatC::cat23C cassette and/or the amiC9 mutation (Fig. 6A) to accept codY inactivation.
Strain (R246) and its amiC9 (R3003), fatC::cat23C (R3002) and amiC9 fatC::cat23C (R3004) derivatives were used as recipients for the codY::trim or codY::spc cassettes, respectively, carried on TD80 and TD81 chromosomal DNA. These recipients contained a hexA- mutation, which impairs mismatch repair, first to ensure similar integration frequency of the cassettes since transfer of the trim cassette is otherwise predicted to be reduced by the Hex system due to the presence of flanking mismatches (see above); second to equalize transformation frequencies of amiCG1459T and fatC C496T point mutations.
The amiC fatC double mutant derivative readily accepted the codY (spc or trim) cassette with the expected frequency relative to the reference marker str41 (∼0.20), while introduction of the cassette into its wild-type parent occurred with a ∼45-fold reduced frequency (Fig. 6B). As expected, amiC and fatC single mutants displayed a ∼5- to 10-fold reduction in transformation frequency of the codY cassette compared with amiC fatC cells. The latter reduction reflected the need for co-transformation of the fatC::cat23C and amiC9 mutation, respectively, while co-transformation of both amiC and fatC together with the codY cassette was required in wild-type cells. Altogether, these results demonstrated that the amiC fatC combination fully suppressed the inviability of codY mutant cells.
Growth and spontaneous competence development of wild type (TD82) and amiC9 (TD84), fatC::cat23C (TD138), amiC9 fatC::cat23C (TD135) and amiC9 fatC::cat23CcodY::trim (TD154) mutant strains were then compared (Fig. S5). Both fatC and amiC mutant strains displayed a cup phenotype, consistent in the latter case with the phenotype of obl mutants (Alloing et al., 1998). The cup phenotype of the double mutant was most similar to that observed with the amiC mutant (Fig. S5) and with the socY strain (Fig. 5E and F). Finally, while codY inactivation resulted in slower growth confirming the observation in Fig. 5, it did not significantly attenuate the cup phenotype observed with the amiC fatC double mutant (compare TD135 and TD154, Fig. S5).
- Top of page
- Experimental procedures
- Supporting Information
CodY is a pleiotropic regulator in low-G+C Gram-positive bacteria involved in the control of different processes like aa uptake, competence, sporulation and virulence. We first established that codY is an essential gene in the D39 strain of S. pneumoniae and in several of its laboratory derivatives by using different genetic approaches (mariner mutagenesis, complementation, transfer of mutations in different genetic backgrounds; Figs 1–3). We then showed that a previously obtained codY mutant (D39ΔcodY) used to identify the CodY regulon (Hendriksen et al., 2008) contains additional suppressor mutations, called socY, one of which was concluded to correspond to a transition mismatch or a frameshift (Fig. 4). We also showed that both the codY socY and socY strains displayed a competence upregulated or cup phenotype (Fig. 5).
Whole-genome sequencing of strains D39 and D39ΔcodY undertaken to identify the socY suppressors revealed the presence of mutations in fatC and amiC, which encode the membrane permease component of ferric iron/haem and oligopeptide ABC transporters respectively (Fig. 6A). This combination of mutations was unambiguously demonstrated to be required to allow tolerance of codY inactivation (Fig. 6B). These results fully confirm the conclusions from codY- transfer experiments, including the presence of a mutation susceptible to mismatch repair (Fig. 4D), i.e. the fatCC496T transition.
Proposed scenario for the initial stabilization/survival of D39ΔcodY cells
Genome sequence clearly indicated that the fatC mutation arose first. However, genome sequence data also strongly suggested that inactivation of amiC is absolutely required for tolerance of codY inactivation. The finding that three different amiC mutations occurred in the D39ΔcodY strain, as confirmed by individual subclone analysis (Table S2), implied a strong selection pressure (growth advantage) for amiC mutant derivatives. The amiC mutations presumably arose independently in an otherwise codY fatC lineage.
It is of note that the D39ΔcodY strain described in Hendriksen et al. (2008) was obtained by backcrossing D39 using chromosomal DNA isolated from an initial D39ΔcodY::trim transformant. Despite this careful strategy, D39ΔcodY::trim turned out to contain two additional mutations, as shown in this study. While a second round of transformation normally eliminates unlinked mutations, it is obviously unable do so when the unlinked mutations are absolutely required for survival. In addition, we noticed that the two suppressor mutations are not very distant. The two closest mutations (fatCC496T and amiCG1459T), located 17 987 nt apart, could frequently be carried by the same DNA fragment. Careful comparison of transformation frequencies with a well-characterized reference marker (preferably present on the same chromosomal DNA) is thus the only way to establish unambiguously that any construct is well tolerated and does not affect cell viability.
CodY, socY and competence
It was implicit in our working hypothesis connecting competence regulation to oligopeptide uptake via a global regulator (like CodY) sensing aa pools (see Introduction) that this regulator would act as a repressor of competence (Claverys et al., 2000). Our data rather suggest that, in contrast to this expectation, inactivation of codY has no major effect on spontaneous competence, most of the effects seen being attributable to the socY mutations (Figs 5 and S5). In fact, the amiC mutation is in itself sufficient to account for the strong derepression of competence development observed in D39ΔcodY cells (Fig. S5). It is of note that despite the cup phenotype documented in this study for the D39ΔcpsΔcodY strain (Fig. 5), previous transcriptome analysis did not reveal induction of any com gene (Hendriksen et al., 2008); presumably, culture conditions used for mRNA extraction did not lead to differential induction of the com regulon between the codY mutant and its parent, competence being either similarly induced or repressed for both strains. While normal Ami functioning somehow leads to competence repression, CodY either is neutral (Fig. S5) or possibly acts as an activator of competence, the latter conclusion being suggested by the attenuation of the cup phenotype of amiC fatC cells upon introduction of the codY knock-out (Fig. 5). If confirmed, the role of S. pneumoniae CodY with respect to competence regulation would thus differ from that of its B. subtilis orthologue, which represses competence (Serror and Sonenshein, 1996), suggesting that the impact of nutrient deprivation on competence induction is opposite in B. subtilis and S. pneumoniae or at least that nutritional signals are conveyed in a very different way in these two species. However, since the impact of codY inactivation on competence could so far be evaluated only in complex genetic backgrounds (i.e. mutant for both amiC and fatC), further work using different approaches, such as a transient depletion of CodY, would be necessary to evaluate the exact role of CodY in the regulation of pneumococcal competence.
CodY, socY and previous transcriptome studies
The finding that the D39ΔcodY strain used in a previous study (Hendriksen et al., 2008) was in fact a codY amiC fatC triple mutant raises the question of the respective contribution of each mutation to the phenotypes previously attributed solely to the inactivation of codY. As concerns transcriptome data, upregulation of aliA (aliB was also upregulated but only 1.8-fold), amiA–amiC–amiD and fatD–fatC–fecE–fatB was reported in ΔcodY (Hendriksen et al., 2008). Binding of CodY to PamiA (as well as to PaliB) was consistent with a direct regulation of oligopeptide uptake by CodY. CodY thus exhibited a strong affinity for PamiA although the effect of BCAA addition was limited (1.56-fold versus 2- to 20-fold enhancement for other promoters). On the other hand, the Kds (the CodY concentration at which 50% of the fragment is shifted) for the PfatD promoter was higher than the highest concentration tested (2000 nM) indicative of a rather weak affinity and was unaffected by addition of BCAAs, contrary to most other CodY-regulated promoters (Hendriksen et al., 2008). Possibly, the binding of CodY to PfatD is affected in the presence of other regulators (see below), as fat regulation is clearly multifactorial (Ulijasz et al., 2009). Alternatively, since ΔcodY cells used to prepare mRNA for transcriptome analysis were also mutant for fatC (i.e. they lacked the main iron transporter of S. pneumoniae), the upregulation of the fatD–fatC–fecE–fatB operon could be a reflection of (inefficient) attempts at iron homeostasis by increasing expression of this transporter rather than the consequence of a lack of repression in the absence of CodY. The same reasoning may apply to dpr gene expression. This gene (spd_1402 in D39; spr1430 in R6; sp1572 in TIGR4) encodes a conserved iron storage-peroxide resistance protein (Ulijasz et al., 2004) that is possibly essential (Pericone et al., 2003). It was found to be downregulated in the codY mutant (Hendriksen et al., 2008). Depletion of iron in codY mutant cells resulting from fatC inactivation could possibly account for the observed reduction in dpr expression without implying any direct regulation by CodY. This would be consistent with the failure to identify a sequence resembling the CodY box upstream of the dpr gene. Clearly, additional work is necessary to establish whether CodY regulates dpr expression directly or indirectly.
CodY, socY and previous virulence studies
Similarly, the presence of amiC and fatC mutations in the D39ΔcodY strain used in a previous study (Hendriksen et al., 2008) raises the question of their respective contribution to the virulence phenotypes previously attributed solely to codY inactivation. The importance of iron for bacterial growth and virulence is well established. Thus, signature-tagged mutagenesis (STM) screens for genes essential during pneumococcal pneumonia identified insertions in genes belonging to each of the three iron transporters characterized in S. pneumoniae (Brown et al., 2001; 2002), piuB (i.e. fatD), pitB (pitADBC operon) and piaA (piaBCD operon) (Hava and Camilli, 2002). These iron uptake systems have been demonstrated to be important for full pneumococcal virulence, especially upon simultaneous mutation of two operons (Brown et al., 2001; 2002). Single mutation of piuB resulted in only a mild reduction of virulence in a pneumonia model, while no attenuated phenotype was observed in systemic infection (Brown et al., 2001).
Several studies have indicated a role for the Ami-AliA/AliB permease, encoded by the amiACDEF operon and the aliA and aliB genes, in pneumococcal virulence. Mutants in amiA and amiC displayed diminished adherence to pulmonary epithelial cells in vitro (Cundell et al., 1995), and aliB and amiACD were identified in STM screens of pneumococcal pneumonia in a serotype 3 and 4 background respectively (Lau et al., 2001; Hava and Camilli, 2002). Furthermore, using a collection of aliA, aliB and amiA single or triple mutants, the Ami-AliA/AliB permease was shown to be required for successful nasopharyngeal colonization, but not for pneumococcal pneumonia, with the most pronounced phenotype for the triple and the aliA and amiA single mutants (Kerr et al., 2004).
Inactivation of codY was previously reported to result in reduced adherence to nasopharyngeal cells and reduced colonization in a mouse model of pneumococcal infection (Hendriksen et al., 2008). However, no significant differences in bacterial loads between wild type and the codY mutant were observed in pneumonia and bacteraemia models of infection. In light of the above, we consider it unlikely that the truncation of fatC contributed to the virulence phenotype associated with codY inactivation. On the other hand, mutation of amiC may, at least partially, have been responsible for the observed attenuation during colonization, but further experiments are needed to exactly determine the impact of the amiC mutation on the virulence phenotype of the codY mutant.
Why is CodY essential in S. pneumoniae?
The finding that CodY is essential suggests that this regulator controls genes/functions that are crucial for S. pneumoniae. The demonstration that mutations in fatC and amiC fully restore viability of codY mutant cells may provide some clue as to these functions. As the genome sequence analysis of D39ΔcodY strongly suggested that the fatC mutation arose first during establishment of the codY mutant construct, it is tempting to speculate about a possible toxicity of iron in cells lacking CodY. If CodY is a repressor of the fatD–fatC–fecE–fatB operon as previously concluded (Hendriksen et al., 2008), derepression of iron uptake in its absence may lead to accumulation of toxic concentrations of iron within the cell, thus creating a strong selection pressure for the emergence of fat mutants. In this context, what could account for the concomitant accumulation of amiC mutations? Considering only Ami and CodY, it could be speculated that codY inactivation leading to derepression of the many transporters of aa controlled by CodY and of Ami results in lethal imbalance of amino acid pools. However, this explanation is not readily connected to the co-occurrence of fatC inactivation. In search of a possible direct connection, we came across the observation that the housekeeping dipeptide permease of Escherichia coli allows utilization of haem as an iron source (Létofféet al., 2006). Haem utilization has thus been shown to require a permease made up of DppBCDF (dipeptide inner membrane transporter) as the ABC transporter and either MppA (periplasmic l-alanyl-γ-d-glutamyl-meso-diaminopimelate-binding protein) or DppA (periplasmic dipeptide-binding protein) as the substrate-binding protein. As S. pneumoniae is devoid of a Dpp system, it is tempting to speculate that the Ami-Obl oligopeptide transporter homologous to Dpp could allow haem utilization as an iron source. The simultaneous involvement of CodY in ami-obl and fat-fec repression, and in dpr activation would, upon codY inactivation, generate a severe oxidative stress because of the simultaneous derepression of iron uptake via two transporters and depletion of the iron storage-peroxide resistance Dpr protein (Fig. 7), hence the observed accumulation of fat and ami mutations in the ΔcodY background. It may not be a mere coincidence that the ami and fat mutations which restore viability of codY mutant cells lead to competence (X-state) derepression (Figs 5 and S5). Since X-state is considered a pneumococcal substitute for SOS (Claverys et al., 2006; Prudhomme et al., 2006), its induction may counteract oxidative stress in codY mutant cells. Further investigation should reveal whether induction of the X-state is required for tolerance of codY inactivation.
Transcriptome analysis revealed a striking parallel between CodY and RitR, an orphan two-component signal transduction response regulator. Increased transcription of piuB and piuA (i.e. fatD and fatB), as well as of amiC and decreased expression of dpr were reported in ritR mutant cells (Ulijasz et al., 2004) suggesting that CodY and RitR share these targets (Fig. 7). RitR was shown to bind three sites (RRB1–3; Fig. 6A) in the promoter region of the fat operon (Ulijasz et al., 2004). Recently, regulation of RitR binding at the fat promoter by a Ser–Thr kinase-phosphatase, StkP–PhpP, was documented (Ulijasz et al., 2009). While RitR and PhpP jointly participate in complex formation at the fat promoter in vitro, addition of StkP was shown to disrupt the complexes (Ulijasz et al., 2009). This finding was consistent with DNA microarray analyses of transcripts from an stkP knock-out showing dependence of fat expression on StkP (Saskova et al., 2007). The overlap between the CodY box (AATTGTCAGAAATT located three nucleotides upstream of the −35 promoter box) and the first RitR box (Fig. 6A) suggests that CodY may also interfere with RitR binding and adds a degree to the complexity of fat regulation. CodY may thus represent one of the additional fat regulators, the existence of which was proposed to account for the failure to detect the opposite effects predicted for the individual deletion of phpP and stkP (Ulijasz et al., 2009). In line with the oxidative stress hypothesis (Fig. 7), ritR mutant cells were shown to display greatly increased susceptibility to streptonigrin, which requires the presence of intracellular iron, as well as to hydrogen peroxide (Ulijasz et al., 2004). Iron overload due to derepression of fat could be responsible for the latter by increasing intracellular free iron concentration and therefore the potential for synthesis of reactive oxygen intermediates. If both CodY and RitR are important for iron homeostasis and the reason for CodY essentiality is to prevent oxidative stress, what about the viability of ritR mutant cells? As there was no mention of difficulty in generating and/or growing ritR mutants, CodY might be more important for ami, fat and dpr regulation than RitR. Alternatively, RitR and CodY could be equally important and the problematic viability of ritR mutant cells has been overlooked. It would therefore be interesting to check whether a previously constructed ritR mutant is readily transferred by transformation.
Further work with other clinical isolates of S. pneumoniae is necessary to establish whether the essentiality of CodY is a general feature of this species. To the best of our knowledge, this is the first report in any bacterium that a member of the CodY family is essential. Is this situation unique to S. pneumoniae? It would be interesting to investigate whether CodY is essential in species closely related to S. pneumoniae and in streptococci in general. It is possible that CodY is also essential in other species but that the presence of suppressor mutation(s) has been overlooked. Our observations may thus prompt careful re-examination of the viability of the codY mutants previously constructed in other species. More generally, the above described tests of acceptance frequencies by transformation should be used routinely when working with ‘important’ genes, to prevent the presence of suppressors going undetected in the future. In any case, because of its essentiality in a major human pathogen, CodY constitutes a potentially interesting new therapeutic target.
- Top of page
- Experimental procedures
- Supporting Information
Bacterial strains, culture and transformation conditions
Streptococcus pneumoniae strains and plasmids used in this study are described in Table 1. Stock cultures were routinely grown at 37°C in Todd–Hewitt plus yeast extract (THY) medium to OD550 = 0.3; after addition of 15% glycerol, stocks were kept frozen at −70°C. To investigate spontaneous competence development, cells were gently thawed and aliquots were inoculated (1 into 25) in C+Y. The initial pH value was adjusted to 7.0 and trypsin (2 µg ml−1) was added to prevent spontaneous competence induction in the pre-culture. After incubation at 37°C to OD550 = 0.2, cultures were centrifuged and cells were concentrated to OD550 of 0.4 in fresh medium containing 15% glycerol and kept frozen at −70°C. For the monitoring of growth and spontaneous competence development, these pre-cultures were gently thawed and aliquots were inoculated (1 into 50, unless otherwise indicated) in luciferin-containing C+Y medium and distributed into a 96-well microplate (300 µl per well). Measurement of competence involved the use of an ssbB::luc transcriptional fusion which reports on competence through light emission by luciferase (Prudhomme and Claverys, 2007). RLU (relative luminescence unit) and OD values were recorded throughout incubation at 37°C (in a Varioskan Flash luminometer; Thermo Electron Corporation).
CSP-induced transformation was performed as described previously (Martin et al., 2000), using precompetent cells treated at 37°C for 10 min with synthetic CSP1 (100 ng ml−1). After addition of transforming DNA, cells were incubated for 20 min at 30°C. Transformants were selected by plating on CAT agar supplemented with 4% horse blood, followed by challenge with a 10 ml overlay containing chloramphenicol (4.5 µg ml−1), erythromycin (0.05 µg ml−1), kanamycin (250 µg ml−1), methotrexate (2.2 µg ml−1), spectinomycin (100 µg ml−1), streptomycin (200 µg ml−1) or trimethoprim (20 µg ml−1), after phenotypic expression for 120 min at 37°C.
Mutagenesis and duplication of codY
Insertions of kan (KanR) or spc (SpcR) minitransposons were generated by in vitro mariner mutagenesis as described (Prudhomme et al., 2007). Plasmids used as a source for the minitransposons were pR410 and pR412 respectively (Table 1). Briefly, plasmid DNA (∼1 µg) was incubated with a target PCR fragment (indicated in the legend of Fig. 1) in the presence of purified Himar1 transposase, leading to random insertion of the minitransposon within the fragment. Gaps in transposition products were repaired as described (Prudhomme et al., 2007) and the resulting in vitro generated transposon insertion library was used to transform S. pneumoniae. Location and orientation of minitransposon insertions were determined as previously described (Prudhomme et al., 2007) through PCR reactions using primers MP127 or MP128 in combination with either one of the primers used to generate codY PCR fragments (Table 1). Cassette–chromosome junctions were sequenced for some insertions as indicated in the legend of Fig. 1.
Placement of a second copy of codY under the control of the maltose-inducible PM promoter at CEP was achieved by cloning into NcoI–BamHI-digested pCEP2 plasmid DNA a codY PCR fragment generated using the codYatg–codYstop primer pair (Table 1) and digested with BamHI and NdeI. The resulting recombinant plasmid pCEP–codY was used as donor in transformation of strain R1501 followed by selection for a KanR transformant, thus generating strain R2349 (Table 1). Plasmid pCEP2 was generated in this study as a high-copy-number derivative of plasmid pCEP (Guiral et al., 2006). Briefly, an EcoRI–PstI fragment from pCEP was ligated to EcoRI–PstI-digested pKL147 (Table 1) to replace the pSC101 replication machinery of pCEP by the pBR replication machinery and ApR resistance gene of pKL147.
Reversion of codY knock-out by transformation
To replace codY::trim or codY::spc insertions by codY+, we took advantage of the kan90Cmariner insertion. This insertion is located immediately upstream the CodY binding site (CYB) in the codY promoter region (Fig. 1A) and does not inactivate codY; we used it as a marker to select for the re-introduction of the codY+ gene by co-transformation with KanR.
To allow the use of a similar strategy for replacement of codY::trim in kanR strains (such as TD83), the kan90C cassette was exchanged with the spcR cassette by transformation of strain R2641 with plasmid pR412 DNA, selecting for SpcR transformants to generate strain R2644 (genotype referred to as spc90C::codY+; Table 1). The exchange is based on the presence of DNA homology at the borders of the synthetic spc and kan minitransposons, allowing exchange of the resistance cassette genes by homologous recombination during transformation.
Whole-genome sequencing of D39 and D39ΔcodY
Roche 454 FLX whole-genome sequencing was performed by Agowa Genomics (Berlin, Germany) using genomic DNA isolated from mid-log cultures by the Genomic DNA kit (Qiagen). For each strain, a shotgun library and a 3 kb span paired end library were generated according to Roche standard protocols, mixed in equal parts (about 400 000 beads from each library) and sequenced using default settings on a one-fourth picotitre plate. A total of 240 496 reads of which 65 336 contained paired ends were obtained for D39ΔcodY (29-fold coverage), and 210 631 reads with 60 676 paired ends were obtained for D39 (25-fold coverage). De novo assembly was carried out using the Roche 454 Newbler software [Release 2.3 (091027_1459)], resulting in 68 contigs in five scaffolds for D39ΔcodY, and 85 contigs in five scaffolds for D39.
Data from the sequencing runs were mapped to the reference D39 strain (Accession No. NC_008533.1) and the variations thereto scored using the Roche 454 Reference Mapper software [Release 2.3 (091027_1459)].
Sequencing of amiC and fatC of individual D39ΔcodY clones
The amiC and fatC sequence of the D39ΔcodY population was verified by a PCR-sequencing approach. To this end, chromosomal DNA was isolated from individual clones by cetyl-trimethylammonium bromide (CTAB) extraction as described previously (van Soolingen et al., 1994). The amiC and fatC loci were PCR-amplified under standard conditions using, respectively, primers HBDamiCF1 and HBDamiCR2 and HBDfatCF and HBDfatCR (Table 1). Subsequently, both strands were sequenced using the same primers used for PCR as well as internal primers HBDamiCR1 and HBDamiCF2 (Table 1) in case of amiC.
- Top of page
- Experimental procedures
- Supporting Information
We thank Nathalie Campo and Calum Johnston for critical reading of the manuscript. We also wish to thank Chantal Granadel for expert technical assistance. This work was supported in part by the European Community's Seventh Framework Programme FP7/2007-2013 under Grant Agreement No. HEALTH-F3-2009-222983 (Pneumopath project). S.C. was the recipient of a PhD thesis fellowship from the Ministère de la Recherche (2007–2009) and from the Fondation pour la Recherche Médicale (2010).
- Top of page
- Experimental procedures
- Supporting Information
- 1998) Systematic identification of essential genes by in vitro mariner mutagenesis. Proc Natl Acad Sci USA 95: 8927–8932. , , , , , and (
- 1994) Three highly homologous membrane-bound lipoproteins participate in oligopeptide transport by the Ami system of the Gram-positive Streptococcus pneumoniae. J Mol Biol 241: 44–58. , , and (
- 1998) Development of competence in Streptococcus pneumoniae: pheromone auto-induction and control of quorum-sensing by the oligopeptide permease. Mol Microbiol 29: 75–84. , , , and (
- 2008) RadC, a misleading name? J Bacteriol 190: 5729–5732. , , , and (
- 2002) Uptake of transforming DNA in Gram-positive bacteria: a view from Streptococcus pneumoniae. Mol Microbiol 45: 411–421. , , , , and (
- 2007) Analysis of the in vitro transcriptional response of human pharyngeal epithelial cells to adherent Streptococcus pneumoniae: evidence for a distinct response to encapsulated strains. Infect Immun 75: 5489–5499. , , and (
- 2001) A Streptococcus pneumoniae pathogenicity island encoding an ABC transporter involved in iron uptake and virulence. Mol Microbiol 40: 572–585. , , and (
- 2002) Characterization of pit, a Streptococcus pneumoniae iron uptake ABC transporter. Infect Immun 70: 4389–4398. , , , and (
- 2007) Search for genes essential for pneumococcal transformation: the RADA DNA repair protein plays a role in genomic recombination of donor DNA. J Bacteriol 189: 6540–6550. , , , , , , and (
- 2001) Regulation of Streptococcus pneumoniae clp genes and their role in competence development and stress survival. J Bacteriol 183: 7295–7307. , , , and (
- 1987) Modulation of competence for genetic transformation in Streptococcus pneumoniae. J Gen Microbiol 133: 1959–1967. , and (
- 2002) Extra-cellular peptide control of competence for genetic transformation in Streptococcus pneumoniae. Front Biosci 7: 1798–1814. , and (
- 1986) Heteroduplex deoxyribonucleic acid base mismatch repair in bacteria. Microbiol Rev 50: 133–165. , and (
- 2000) Is the Ami-AliAB oligopeptide permease of Streptococcus pneumoniae involved in sensing environmental conditions? Res Microbiol 151: 1–7. , , and (
- 2006) Induction of competence regulons as general stress responses in Gram-positive bacteria. Annu Rev Microbiol 60: 451–475. , , and (
- 1995) Peptide permeases from Streptococcus pneumoniae affect adherence to eucaryotic cells. Infect Immun 63: 2493–2498. , , , , and (
- 2004) Interconnection of competence, stress and CiaR regulons in Streptococcus pneumoniae: competence triggers stationary phase autolysis of ciaR mutant cells. Mol Microbiol 51: 1071–1086. , , , , , , et al. (
- 1989) Repair of single- and multiple-substitution mismatches during recombination in Streptococcus pneumoniae. Genetics 121: 29–36. , , and (
- 2001) Pleitropic transcriptional regulator CodY senses the intracellular pool of branched-chain amino acids in Lactococcus lactis. Mol Microbiol 40: 1227–1239. , , , , and (
- 2005) Competence-programmed predation of noncompetent cells in the human pathogen Streptococcus pneumoniae: genetic requirements. Proc Natl Acad Sci USA 102: 8710–8715. , , , and (
- 2006) Construction and evaluation of a chromosomal expression platform (CEP) for ectopic, maltose-driven gene expression in Streptococcus pneumoniae. Microbiology 152 (Special Issue on Pneumococcus): 343–349. , , , , , , and (
- 2008) Interaction of Bacillus subtilis CodY with GTP. J Bacteriol 190: 798–806. , , and (
- 2002) Large-scale identification of serotype 4 Streptococcus pneumoniae virulence factors. Mol Microbiol 45: 1389–1406. , and (
- 2006) New insights into pneumococcal fratricide: relationship to clumping and identification of a novel immunity factor. Mol Microbiol 59: 1297–1307. , , , , and (
- 2008) CodY of Streptococcus pneumoniae: link between nutritional gene regulation and virulence. J Bacteriol 190: 590–601. , , , , , , et al. (
- 2009) Pneumococcal LytR, a protein from the LytR-CpsA-Psr family, is essential for normal septum formation in Streptococcus pneumoniae. J Bacteriol 191: 5859–5864. , and (
- 2004) The Ami-AliA/B permease of Streptococcus pneumoniae is involved in nasopharyngeal colonization but not in invasive disease. Infect Immun 72: 3902–3906. , , , , , , et al. (
- 2007) Genome sequence of Avery's virulent serotype 2 strain D39 of Streptococcus pneumoniae and comparison with that of unencapsulated laboratory strain R6. J Bacteriol 189: 38–51. , , , , , , et al. (
- 2001) A functional genomic analysis of type 3 Streptococcus pneumoniae virulence. Mol Microbiol 40: 555–571. , , , , , , et al. (
- 1979) Donor deoxyribonucleic acid length and marker effect in pneumococcal transformation. J Bacteriol 138: 80–86. , , and (
- 1998) Localization of bacterial DNA polymerase: evidence for a factory model of replication. Science 282: 1516–1519. , and (
- 2006) The housekeeping dipeptide permease is the Escherichia coli heme transporter and functions with two optional peptide binding proteins. Proc Natl Acad Sci USA 103: 12891–12896. , , and (
- 2000) Cross-regulation of competence pheromone production and export in the early control of transformation in Streptococcus pneumoniae. Mol Microbiol 38: 867–878. , , , , and (
- 2003) Additional targets of the Bacillus subtilis global regulator CodY identified by chromatin immunoprecipitation and genome-wide transcript analysis. J Bacteriol 185: 1911–1922. , , , , , , and (
- 1998) Competence-specific induction of recA is required for full recombination proficiency during transformation in Streptococcus pneumoniae. Mol Microbiol 27: 159–170. , , , and (
- 2003) Factors contributing to hydrogen peroxide resistance in Streptococcus pneumoniae include pyruvate oxidase (SpxB) and avoidance of the toxic effects of the fenton reaction. J Bacteriol 185: 6815–6825. , , , and (
- 2007) There will be a light: the use of luc transcriptional fusions in living pneumococcal cells. In The Molecular Biology of Streptococci. Hakenbeck, R., and Chhatwal, G.S. (eds). Norfolk, UK: Horizon Scientific Press, pp. 519–524. , and (
- 2006) Antibiotic stress induces genetic transformability in the human pathogen Streptococcus pneumoniae. Science 313: 89–92. , , , , and (
- 2007) In vitro mariner mutagenesis of Streptococcus pneumoniae: tools and traps. In The Molecular Biology of Streptococci. Hakenbeck, R., and Chhatwal, G.S. (eds). Norfolk, UK: Horizon Scientific Press, pp. 511–518. , , and (
- 1993) Characterization of the Streptococcus pneumoniae maltosaccharide regulator MalR, a member of the LacI-GalR family of repressors displaying distinctive genetic features. J Biol Chem 268: 25402–25408. , , and (
- 2001) Bacillus subtilis CodY represses early-stationary-phase genes by sensing GTP levels. Genes Dev 15: 1093–1103. , , , and (
- 2007) Eukaryotic-type serine/threonine protein kinase StkP is a global regulator of gene expression in Streptococcus pneumoniae. J Bacteriol 189: 4168–4179. , , , and (
- 1996) CodY is required for nutritional repression of Bacillus subtilis genetic competence. J Bacteriol 178: 5910–5915. , and (
- 2005) CodY, a global regulator of stationary phase and virulence in Gram-positive bacteria. Curr Opin Microbiol 8: 203–207. (
- 2007) Control of key metabolic intersections in Bacillus subtilis. Nat Rev Microbiol 5: 917–927. (
- 1994) DNA fingerprinting of Mycobacterium tuberculosis. Methods Enzymol 235: 196–205. , , , and (
- 2001) An rpsL cassette, Janus, for gene replacement through negative selection in Streptococcus pneumoniae. Appl Environ Microbiol 67: 5190–5196. , , , and (
- 2003) A solute binding protein of Streptococcus pneumoniae iron transport. FEMS Microbiol Lett 220: 303–308. , , and (
- 2002) Identification of 113 conserved essential genes using a high-throughput gene disruption system in Streptococcus pneumoniae. Nucleic Acids Res 30: 3152–3162. , , , , and (
- 1975) Marker discrimination in deoxyribonucleic acid-mediated transformation of various pneumococcus strains. J Bacteriol 121: 608–618. , , and (
- 2004) Regulation of iron transport in Streptococcus pneumoniae by RitR, an orphan response regulator. J Bacteriol 186: 8123–8136. , , , and (
- 2009) Phosphorylation of the RitR DNA-binding domain by a Ser-Thr phosphokinase: implications for global gene regulation in the streptococci. Mol Microbiol 71: 382–390. , , and (
- Top of page
- Experimental procedures
- Supporting Information
|MMI_7339_sm_FigS1-5_and_TableS1-2.pdf||1397K||Supporting info item|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.