Moonlighting proteins activate transformation in epigenetically-differentiated phase variants of multidrug-resistant Streptococcus pneumoniae

Abstract

Transformation of Streptococcus pneumoniae GPSC1 has enabled vaccine evasion and the acquisition of antibiotic resistance. Epigenetic phase variants of GPSC1 isolate RMV7, differentiated by rearrangements at the tvr restriction-modification locus, differed ∼100-fold in their transformation efficiency. This variation was recapitulated in knock-in mutants of the relevant tvr alleles. RNA-seq showed the difference was due to blocking of the early competence regulatory cascade. The more transformable variant upregulated expression of manLMN, encoding a carbon source importer. This was shown to be necessary for efficient competence induction, despite being dispensable for growth in rich media. Transformation was promoted by import of N-acetylglucosamine, which activated competence through an orthologue of the gram-negative competence regulator TfoX, and an enzyme likely involved in nucleotide-mediated signalling. A mobile genetic element was more active in the less transformable variant, which limited competence induction through inhibiting CIRCE motif binding by the chaperone regulator HrcA. Correspondingly, both heat shock and decreased Ca²⁺ concentrations reduced competence by limiting HrcA binding regulatory sequence motifs. Hence both ManLMN and HrcA moonlighted as activators of competence. Such proteins may be important in potentiating horizontal DNA exchange, as selection on their primary function likely constrains them from mutating into alleles that selfishly downregulate transformation.

Competence for natural transformation was first identified in Streptococcus pneumoniae (the pneumococcus) in the early 20^th century [1]. Cells can be “transformed” to express a new phenotype through the acquisition of exogenous DNA, integrated into their genome through homologous recombination following its import from the environment through the specialised cell-encoded competence machinery [2]. Transformation has played a key role in the emergence of antibiotic-resistant S. pneumoniae, both through generating “mosaic” alleles of core loci [3–5] and the acquisition of specialised resistance genes [6]. It has also enabled vaccine evasion through recombinations affecting the capsule polysaccharide synthesis (cps) locus altering surface antigens [7, 8].

Despite the ability of transformation to accelerate such adaptive evolution in S. pneumoniae, considerable variation in the rate of diversification of strains through this mechanism persists across the species [9]. Epidemiological studies have found the r/m ratio of base substitutions introduced through homologous recombination, relative to those occurring through point mutation, varies from well over 10 [7, 9] to below 0.1 [9, 10] across the species. Similarly, in vitro assays have identified >100-fold differences in the transformation efficiency of S. pneumoniae genotypes, with substantial variation even between isolates of the same serotype or strain [11–14]. Many isolates are routinely found not to be transformable under standard conditions [11]. This is often the consequence of integrative mobile genetic elements (MGEs) disrupting genes necessary for transformation [6,15–17], selfishly preventing themselves from being eliminated from the chromosome [17]. Yet in other non-transformable isolates the highly-conserved competence machinery is intact [11, 18]. This suggests the variation in transformation rates also reflects differences in regulation of the competence system.

The best-characterised signal inducing transformation in S. pneumoniae is the competence stimulating peptide (CSP) pheromone, which acts as a quorum-sensing system that signals through the ComDE two-component regulator [19]. This activates early competence genes after about ten minutes [20]. These include comX, encoding an alternative sigma factor [21]. ComX enables the RNA polymerase to recognise late competence genes [20], which feature a “combox” signal in their promoters [22, 23]. This results in pneumococci entering a transient competent state around 20 minutes post-CSP induction, after which the relevant machinery is degraded [24], and the cells become temporarily refractory to induction [21].

Transformation efficiency is known to vary between isogenic pneumococci through phase variation in capsule production. Transparent colony variants produce less capsule, and are less virulent, but more transformable, than opaque colony variants [25]. This short-term variation has been linked to rapid changes at the phase-variable inverting variable restriction (ivr) locus, encoding the conserved Type I SpnIII restriction-modification system (RMS) and the IvrR recombinase that drives sequence inversions within the locus [26–31]. These rearrangements alter the target recognition domains (TRDs) within the active HsdS specificity protein, which determines the DNA motif that is targeted by both the methylase and endonuclease activities of the system. Consequently, changes at this single locus affect genome-wide methylation patterns, causing pleiotropy through its impact on the transcription of multiple genes [27]. These phase-variable RMSs can thereby maintain phenotypic heterogeneity within a genetically near-homogenous population [32], resulting in “bet hedging” that can increase the chances of a species surviving a changing environment [33, 34].

The second pneumococcal phase-variable Type I restriction-modification system (named SpnIV), encoded by the translocating variable restriction (tvr) locus, varies through TvrR recombinase-mediated excision-reintegration [27,28,35,36]. This locus is generally conserved across pneumococci, but the complement of TRDs varies between strains, increasing the range of HsdS proteins, and possible methylation patterns, across the species [28, 35]. This locus is inactive in the R6x laboratory strain that is typically used to study pneumococcal competence [28,35,37]. Here, we test pairs of otherwise-isogenic tvr variants to identify the regulatory mechanisms underlying phenotypic heterogeneity in genotypes commonly causing pneumococcal disease.

Results

Phase variants of S. pneumoniae RMV7 differ in their transformation efficiency

Four pairs of otherwise-isogenic “locked” phase variants, in which tvrR was knocked out or disrupted by a selectable and counter-selectable Janus cassette marker [35, 38], were screened for differences in their transformation efficiency (Fig. 1A). In the RMV6 and RMV7 pairs, the variant in which the active tvr locus HsdS comprised the TRDs III-i (directing the SpnIV system to target the bipartite motif TGAN₇TATC) was found to have significantly higher transformation efficiency following induction by exogenous CSP. These closely-related genotypes both originate from the multidrug-resistant GPSC1 strain [39]. Many replicates were required to robustly establish the difference between the RMV6 variants, whereas the RMV7 variants exhibited a clearer ∼100-fold difference in their transformation rate (Fig. 1A). This was because the RMV7 variant carrying the alternative form of the tvr locus, with an active HsdS comprising the TRDs III-iii (which directs the SpnIV system to target the motif TGAN₇TCC; Fig. 1B and S1) had an almost-undetectable transformation efficiency. Culturing the wild-type isolate (RMV7_wt) over successive days demonstrated large changes in the relative frequency of these variants, although the less-transformable variant (RMV7 tvrR::Janus; henceforth, RMV7_domi, carrying tvr_domi) was typically more common than the more transformable variant (RMV7 ΔtvrR; henceforth, RMV7_rare, carrying tvr_rare; Fig. 1C). A comparison of RMV7_wt and RMV7_domi found them to have similarly low transformation efficiencies, whereas that of RMV7_rare was confirmed to be ∼50-fold higher, consistent with the relative proportions of the variants observed in vitro (Fig. 1D). Therefore, RMV7_domi and RMV7_rare exhibited distinctive phenotypes that correlated with their tvr locus arrangements.

Figure 1

Variation in transformability between locked tvr variants. (A) Violin plot showing the transformation efficiency of four pairs of tvr locus variants constructed from isolates RMV5, RMV6, RMV7 and RMV8. Each individual point represents an independent transformation experiment. The horizontal line within each violin shows the median for each genotype. The brackets indicate the statistical significance of the comparison between variants from the same isolate background, as calculated using a two-tailed Wilcoxon rank sum test. (B) Schematic of the tvr loci from RMV7_wt, RMV7_domi and RMV7_rare to show the genes encoding the methylase (hsdM), endonuclease (hsdR), regulatory system (tvrAT) and recombinase (tvrR). The variants differ in their active hsdS genes, upstream of tvrATR. The RMV7_domi HsdS protein comprises the TRD combination III-iii (recognising motif TGAN₇TCC), whereas that of RMV7_rare contains the TRDs III-I (recognising motif TGAN₇TATC). The black arrows indicate the binding sites of a forward primer, in hsdM, and reverse primers, in hsdS fragments. (C) Line graph showing the ratio of RMV7_domi to RMV7_rare loci in eight-day passages of RMV7_wt. Although conversion to RMV7_rare was detected, the RMV7_domi variant was more common in each of the three replicates at every measured timepoint. (D) Violin plot showing the higher transformation efficiency of RMV_rare relative to RMV7_domi or RMV7_wt. Each individual point represents an independent experiment in which the number of transformants, and number of overall colony-forming units (cfu), was calculated. The horizontal line within each violin shows the median for each genotype. Both mutants were compared with RMV7_wt; the horizontal bracket shows a significant difference, as calculated from a two-tailed Wilcoxon rank sum test. (E) Violin plot showing the adhesion of variants to an abiotic surface, as quantified by OD₅₅₀ after crystal violet staining of cells resisting washing from a microplate well. The RMV7_rare variant showed a significantly greater adhesion after 16 hours incubation at 35 °C, as assessed by a two-tailed Wilcoxon rank sum test. (F) Violin plots showing the transformation efficiency of knock-in mutants during a passage experiment. The tvr loci of RMV7 tvr_domi::Janus and RMV7 tvr_rare::Janus (both containing a tvrR gene interrupted by a Janus cassette) were each introduced into an RMV7 tvr::cat background (Fig. S4). This isogenic pair were separately passaged in liquid cultures over six days in five independent replicates. The number of transformants observed from three transformation assays, conducted each day for both variants, is shown by the individual points’ shapes and colours. The violin plots summarise the median and distribution of these values. The brackets indicate the statistical significance of the comparison between variants from the same day of the passage, as calculated using a two-tailed Wilcoxon rank sum test. Across all panels, significance is coded as: p < 0.05, *; p < 0.01, **; p < 10^-3, ***; p < 10^-4, ****. All p values were subject to a Holm-Bonferroni correction within each panel.

RMV7_rare was also significantly more adhesive to an abiotic surface (Fig. 1E), which can be considered a proxy for biofilm formation [40]. This could be consistent with RMV7_rare being enriched for transparent phase variants. However, microscopy found no clear difference in colony morphology between the variants (Fig. S2). An alternative explanation for the phenotypic differences could be mutations that occurred during genetic manipulation of the isolates [41]. Alignment of the two variants’ assemblies found them to be distinguished by seven non-synonymous single nucleotide polymorphisms and two premature stop codons outside the tvr locus, none of which were within genes known to directly affect the competence machinery (Table S1). Nevertheless, we tested the effect on transformation of mutations in RMV7_rare that were absent from both the RMV7_domi and RMV7_wt sequences. A non-synonymous change in phoB, encoding a phosphate-sensitive response regulator [42], was not found to affect transformation efficiency (Fig. S3). Similarly, a premature stop codon in pstS was observed in RMV7_rare, but knocking out this gene also failed to explain any difference in transformability between the variants (Fig. S3). Hence the differences between RMV7 variants could not be explained by point mutations or alterations in encapsulation.

To test whether the phenotypic differences were causatively associated with variation in the SpnIV RMS, the tvr loci of RMV7_domi and RMV7_rare were introduced into RMV7_wt tvr::cat. This generated the otherwise isogenic knock-in recombinants RMV7 tvr_domi::Janus and RMV7 tvr_rare::Janus, carrying two different locked tvr loci (Fig. S4). The transformability of both was assayed over five replicate six-day passages. This demonstrated a persistent distinction in the transformation efficiency of the two tvr genotypes, which replicated the difference between the original variants (Fig. 1F). Therefore, the RMV7 tvr_rare and tvr_domi alleles were associated with significant differences in transformation efficiency in otherwise isogenic isolates.

Induction of early and late competence genes is blocked in RMV7_domi

To understand how the tvr loci caused a difference in transformation, RNA-seq was used to quantify patterns of transcription in the recombinants RMV7 tvr_domi::Janus and RMV7 tvr_rare::Janus. Samples were taken pre-CSP, 10 minutes post-CSP, and 20 minutes post-CSP from each of three biological replicates (Fig. S4). The 18 RNA samples were sequenced as 200 nt paired-end multiplexed libraries on a single Illumina HiSeq 4000 lane. After alignment to the RMV7_domi genome, analysis of the RNA-seq data found the fragment length distributions (Fig. S5) and gene expression densities (Fig. S6) were consistent across samples (Table S2). Q-Q plots suggested a Benjamini-Hochberg corrected q value of 10^-3 was an appropriate threshold for identifying significant transcriptional variation (Fig. S7-8). This identified 154 genes that significantly differed in their expression between the two genotypes prior to CSP exposure, or between the pre- and post-CSP samples from the same genotype (Fig. 2 and S9-10; Table S3).

Figure 2

Chromosomal distribution of genes exhibiting significant differences in transcription between RNA-seq samples. The outer ring shows the annotation of RMV7_domi (accession code OV904788). Protein coding sequences are represented as black boxes, with the vertical positioning within the ring indicating the strand of the genome on which they are encoded. The next ring inwards shows significant pre-CSP differences in transcription: green genes were more highly expressed in RMV_domi, and blue genes were more highly expressed in RMV_rare. The next ring inwards show significant changes in gene expression 10 minutes post-CSP in RMV7_domi: pink genes were upregulated, and purple genes were downregulated. The third ring inwards shows significant changes in gene expression 10 minutes post-CSP in RMV7_rare: red genes were upregulated, and orange genes were downregulated. The two inner rings repeat this representation for changes in gene expression 20 minutes post-CSP.

Clustering of the datasets found the biggest separation distinguished the post-CSP RMV7 tvr_rare::Janus transcriptomes from the others (Fig. S11), suggesting a major difference in the induction of the competence system between the variants. In RMV7 tvr_rare::Janus, the early competence genes showed elevated expression 10 minutes post-CSP (Fig. S12), with the late competence genes exhibiting more variable patterns of transcription (Fig. S13). Also upregulated was briC, a competence-induced biofilm formation signal [43], and some nucleotide metabolism and transporter genes (purA, tadA, dut, ribF, adeQ; Fig. S14). Another upregulated transporter, encoded by a three gene cluster, has reportedly been induced in response to exposure to CSP [44] and antimicrobial peptides [45], and consequently was named pieABC (peptide-induced exporter; CDSs IONPJBJN_01324-6 in RMV7_domi, corresponding to SP_0785-787 in TIGR4; Fig. S14, Table S3).

By contrast, CSP upregulated just five genes in RMV7 tvr_domi::Janus: the quorum sensing genes comCDE and briC (Fig. S12), and a gene encoding a CsbD stress response protein that was not CSP-responsive in RMV7 tvr_rare::Janus. However, the induction of comCDE was much weaker, and there was no sign of late competence genes being activated by ComX. However, expression of the comX gene itself can be difficult to determine through RNA-seq, owing to the presence of two near-identical paralogues in pneumococcal genomes [46].

As an independent test of these transcriptional differences, qRT-PCR experiments were undertaken on the original RMV7_domi and RMV7_rare variants, and the control genotypes RMV7_wt and RMV7_wt tvr::cat, the latter of which lacked a tvr locus. Transformation assays performed in parallel with these RNA extractions found RMV7_wt tvr::cat was more efficiently transformable than RMV7_wt or RMV7 tvr_domi::Janus, demonstrating the active, dominant form of the locus was repressing transformation (Fig. S4). The qRT-PCR data showed genotypes of both low (RMV7_wt and RMV7_domi) and high (RMV7_rare and RMV7_wt tvr::Janus) transformation efficiency up-regulated the early competence genes comD and comX in response to CSP (Fig. S15). However, the late competence genes comEA and comYC were only significantly upregulated in the highly transformable strains. Hence the difference in transformability between the RMV7 variants was a consequence of late competence gene activation being blocked in RMV7_domi through consequences of tvr_domi expression.

Differences in pre-CSP gene expression

The two pre-CSP samples were compared to identify the cause of this difference in competence induction. Pre-CSP, 33 genes exhibited higher expression in RMV7 tvr_domi::Janus, and 20 genes exhibited higher expression in RMV7 tvr_rare::Janus (Fig. 2; Table S3). These did not include any cps locus genes, which were more highly expressed in RMV7 tvr_rare::Janus (Fig. S16), confirming that the elevated transformation efficiency of this genotype did not reflect an enrichment of transparent phase variants [27, 47]. To test whether these differentially-expressed genes were affected by methylation within regulatory or promoter sites, the distances from protein coding sequence start codons to the nearest methylation site was plotted for both SpnIV motifs (Fig. S17). There was no general relationship between differential expression and proximal methylation, although four cases of variable methylation sites being within 100 bp of a differentially-expressed gene’s start codon were identified. In three cases, the coding sequences close to tvr_domi motifs were within a larger set of co-transcribed genes, and it was unlikely that transcription initiation would be affected by methylation (Fig. S18). However, one tvr_rare motif was within the promoter of the piuABCD operon (Fig. S18), encoding an iron transporter, that was upregulated in RMV7 tvr_rare::Janus (Fig. S14). However, knocking out piuA did not reduce the transformation efficiency of RMV7_rare, suggesting this change was independent of those affecting competence (Fig. S15).

As the induction of transformation has been suggested to represent a stress response [48], the overall distribution of the SpnIV target motifs was analysed, to test whether differing spatial arrangements across the chromosome might affect cell physiology. While the tvr_rare motifs were uniformly distributed, the RMV7_domi motifs were enriched in one segment of the chromosome (Fig. S20). This suggested that the differences between the two variants was not the result of strong effects at a small number of promoters, but instead a broader response to a general change in the distribution of DNA modifications, as suggested for other epigenetic phase variants [27].

Consistent with the methylation pattern of RMV7_domi reducing transformation efficiency through altering cell physiology, many of the differentially-expressed loci identified by RNA-seq encoded regulators. The mgrA gene, encoding a regulator of cell adhesion, was expressed more highly in RMV7_domi (Fig. S21). Consistent with previous observations, this correlated with repression of the rlrA pilus islet [49] and activation of the accBC-yqhY-amaP-csbD cell wall stress operon (IONPJBJN_01032-6) [50] in RMV7_domi. The cell-surface chaperone-encoding gene prsA was also upregulated, as was the chaperone regulator gene hrcA. The groES-groEL and grpE-dnaK-IONPJBJN_02152-dnaJ operons of the hrcA regulon were also more highly expression in RMV7_domi, although this difference only achieved genome-wide significance for IONPJBJN_02152 (Fig. S22; Table S3).

Another indicator of cell stress in RMV7_domi was the increased activity of a phage-related chromosomal island (PRCI; also known as a phage-inducible chromosomal island, or PICI) integrated adjacent to dnaN (PRCI_dnaN; Fig. S23; Table S3). This is one of two PRCIs associated with GPSC1 [7, 28], the other being integrated near uvrA (Fig. 2). The regulatory mechanisms of these elements are not thoroughly characterised [51], and in the absence of a helper prophage, it is unclear exactly what signal may have triggered this increased activity. Given integrative MGEs are likely under selection to reduce host cell transformability [17], the increased activity of PRCI_dnaN could drive inhibition of the competence machinery. A further indicator of cell stress was the increased expression of the ciaRH two-component system, which is known to inhibit competence [52]. Hence PRCI_dnaN and ciaRH were the primary candidates for causing the observed difference in transformability between the RMV7 variants.

ManLMN links competence induction to sugar metabolism

The higher expression of the ciaRH genes in the non-transformable genotypes RMV7_domi and RMV7_wt, relative to the transformable genotype RMV7_rare, was confirmed by qRT-PCR (Fig S15). CiaRH binds at least 15 promoter sequences, five of which drive the expression of cia-dependent small RNAs (csRNA) that suppress the induction of competence for transformation through inhibiting expression of the comC gene, encoding CSP [53]. The remaining ten drive the expression of protein coding sequences (CDSs) [53]. Although csRNA5 was the only non-coding sequence that appeared to respond to ciaRH activation, many of the CiaRH regulon CDSs were more highly expressed in RMV7_domi compared to RMV7_rare pre-CSP (Fig. S24). These included the extracytoplasmic chaperone and serine protease HtrA [53–55], which can block competence induction at low CSP concentrations through degrading extracellular CSP [54, 56]. In agreement with some previous studies, elimination of htrA further increased the transformability of RMV7_rare [54], but the same mutation had no significant effect in RMV7_wt (Fig S25). This suggests HtrA inhibits the induction of competence, but is unlikely to explain much of the difference between these variants. Similarly, knock out of ciaRH itself had little effect in either genotype (Fig. S25), suggesting an alternative locus was causing the difference in transformation efficiency between the variants.

Among the CiaRH regulon, the most significant difference in expression between RMV7_rare and RMV7_domi was the ManLMN carbon source importer operon (Table S3). The manLMN genes were more highly expressed in RMV7_rare, as CiaRH binds promoter sequences within the operon [37,53,57,58] and acts as a repressor [59]. Disrupting the manLMN locus reduced the transformation efficiency of RMV7_rare by >5-fold, while having little effect in RMV7_wt (Fig. 3A-B). To test whether the observed transformation differences reflected a growth defect, manLMN::Janus mutants of both RMV7_wt and RMV7_rare were cultured in the same rich media (Fig. S26). However, removal of manLMN had little effect on growth, suggesting the transporter’s effect on transformation was through regulation rather than proliferation. Hence the lower expression of manLMN in RMV7_domi accounts for some of its reduced transformation efficiency relative to RMV7_rare.

Figure 3

The dependence of transformation efficiency on import of carbon sources. (A) Violin plots showing the transformation efficiency of RMV7_wt relative to a mutant in which manLMN had been disrupted by a Janus cassette. Each genotype was transformed in unsupplemented media, and in the presence of one of six carbon sources, as indicated by the plot colour (see key on bottom right). Each point represents an independent experiment, and the horizontal line within the violin plots show the median for each combination of recipient cell genotype and carbon source. For each genotype, Wilcoxon rank-sum tests were used to test for evidence of changes in transformation efficiency caused by each carbon source. Significant differences are indicated by the black brackets at the top of the panel. (B) Violin plots showing the transformation efficiency of RMV7_rare relative to a mutant in which manLMN had been disrupted by a Janus cassette. Data are displayed as in panel A. (C) Violin plots showing the transformation efficiency of RMV7_rare relative to a mutant in which tfoX had been disrupted by a Janus cassette, and a third genotype in which tfoX had been restored. Data are displayed as in panel A; the data for RMV7_rare are replicated from panel B. (D) Violin plots showing the transformation efficiency of RMV7_rare relative to a mutant in which yjbK had been disrupted by a Janus cassette, and a third genotype in which yjbK had been restored. Data are displayed as in panel A; the data for RMV7_rare are replicated from panel B. (E) Violin plots showing the concentration of 3’,5’-cAMP in samples taken from exponential and stationary phase cultures of E. coli DH5α, and S. pneumoniae RMV7_wt and RMV7_rare genotypes differing in whether yjbK was intact or not. The genotypes are indicated by the colours (see key on bottom left). (F) Violin plots showing the transformation efficiency of RMV7_wt relative to mutants in which either tfoX or yjbK had been disrupted by a Janus cassette. Data are displayed as in panel A. Across all panels, significance is coded as: p < 0.05, *; p < 0.01, **; p < 10^-3, ***; p < 10^-4, ****. All p values were subject to a Holm-Bonferroni correction within each panel.

ManLMN is a phosphotransferase system (PTS) transporter that can facilitate the import of glucose, mannose, galactose, fructose, aminoglucose and N-acetylglucosamine (GlcNAc) [60]. To test whether any effects were dependent on the imported substrate, growth curves and transformation experiments were undertaken in the presence of these carbon source supplements. Few significant differences in growth rate were observed, except a small positive effect of GlcNAc on the growth of RMV7_wt (Fig. S26). However, GlcNAc increased transformation efficiency ∼10-fold in RMV7_wt (Fig. 3A) and ∼2-fold in RMV7_rare (Fig. 3B). In both variants, this effect was dependent upon manLMN. This was confirmed through disrupting, and then restoring, manL alone (Fig. S27). Therefore, uptake of GlcNAc by ManLMN increased the transformability of RMV7.

N-acetylglucosamine activates competence through TfoX and Yjbk

This result suggested two parallels with the regulation of competence in gram-negative bacteria. Competence in Vibrio cholerae is induced GlcNAc disaccharides [61], thought to be generated from degradation of chitin [62]. This is mediated through the Transformation Factor X (TfoX) protein [63, 64]. An orthologue of this protein (also called Sxy) is also central to regulation of competence in Haemophilus influenzae [65], in which it facilitates induction via 3’,5’-cyclic AMP (cAMP) signalling [66, 67]. Intracellular concentrations of 3’,5’-cAMP rise in many proteobacteria when the primary glucose PTS transporter is inactive, as the accumulation of phosphorylated EIIA PTS subunits stimulates adenylate cyclase activity [68]. We investigated whether analogues of either of these pathways existed in S. pneumoniae.

An orthologue of the N terminal domain of TfoX was annotated, but not described, in S. pneumoniae ATCC 700669 [69]. In RMV7, this gene (tfoX_Spn; IONPJBJN_02097) is conserved in the same position, two genes upstream of the comEA competence operon (Fig. S28). The amino acid sequence corresponded to only the N-terminal region of the gram-negative orthologue, and was predicted to form a four-strand beta sheet flanked by alpha helices (Fig. S29). The gene could be both disrupted, and restored, in RMV7_rare. Transformation assays using a panel of sugars demonstrated the elevated induction of competence by GlcNAc in this variant was dependent on TfoX_Spn (Fig. 3C).

A gene encoding a candidate adenylate cyclase, yjbK, was also identified in RMV7 (IONPJBJN_01639). This protein is predicted to have a β-barrel structure, as observed for orthologous enzymes synthesising 3’,5’-cAMP (Fig. S30). The yjbK gene could be both disrupted, and restored, in RMV7_rare. Transformation assays with these genotypes demonstrated the effect of GlcNAc on competence induction in RMV7_rare also depended on YjbK (Fig. 3D). However, no 3’,5’-cAMP signalling pathway is known in Firmicutes [70]. Therefore, an ELISA was used to compare the 3’,5’-cAMP levels in RMV7_wt, RMV7_rare yjbK::Janus and RMV7_rare yjbK restored, relative to Escherichia coli DH5α. High levels of 3’,5’-cAMP were found in fractions collected from both exponential and stationary phases of E. coli. However, the concentrations of 3’,5’-cAMP detected in S. pneumoniae were close to the lower detection threshold, and unaffected by yjbK presence (Fig. 3E). Additionally, exogenous 3’,5’-cAMP had no effect on transformation efficiencies in any RMV7 genotypes (Fig. S31). Therefore, it is unlikely that YjbK’s regulatory role is mediated through 3’,5’-cAMP production.

To test whether TfoX and YjbK had a similar role in mediating the larger response to GlcNAc in RMV7_wt, both were also knocked out in this background. However, the results showed a different pattern (Fig. 3F). The disruption of tfoX had no effect on GlcNAc’s ability to induce competence, whereas RMV7_wt yjbK::Janus failed to activate competence unless GlcNAc or sialic acid were supplemented in the media. These results were consistent with ManLMN affecting competence through at least two pathways, one of which is dependent upon TfoX and YjbK. Disruption of these genes in the laboratory strain R6x was also found to reduce transformation efficiency, although detecting these effects required culturing in a low-sugar chemically-defined medium (see Methods; Fig. S32).

Variation in mobile element activity associated with cell stress

Although ManLMN had a role in regulating the competence machinery, it could not account for the large difference in transformability between RMV7_domi and RMV7_rare alone. MGEs have been found to strongly affect levels of competence in S. pneumoniae and other species [16,71–73]. Therefore we tested the hypothesis that the increased activity of PRCI_dnaN in RMV7_domi may inhibit the competence of the host cell, in order to prevent the MGE being deleted through homologous recombination [17]. A candidate gene (IONPJBJN_00496) with the potential to limit homologous recombination encoded a protein similar to DNA damage-inducible protein D (DinD), which inhibits RecA activity in E. coli [74]. However, the RMV7_wt IONPJBJN_00496::Janus mutant did not show any increase in transformability (Fig. 4A).

Figure 4.

Effect of removing PRCI_dnaN on RMV7_wt. (A) Violin plots showing the number of transformants observed in assays of RMV7_wt mutants in which different parts of PRCI_dnaN were replaced with a Janus cassette. The genotypes are arranged left to right, and coloured black to purple, to represent the increasing proportion of the element that was replaced by the cassette. Mutants removed different combinations of the dinD-like gene (IONPJBJN_00496); the regulatory genes (rep); the replication genes (rep); the integration genes (int), and the att site. The structure of RMV7 PRCI_dnaN meant that replacing the int-rep region also deleted the intervening reg genes. Each point is an independent transformation assay. The violin plot summarises the result for each mutant, with a horizontal line indicating the median. Each mutant was compared against the parental RMV7_wt genotype using a Wilcoxon rank sum test. Black brackets at the top of the plot indicate significant differences in the number of observed transformants. (B) Violin plot quantifying the effect of PRCI_dnaN on transformation efficiency. This comparison of RMV7_wt with a mutant in which the PRCI and its att site were removed was independent of the experiments presented in panel A, and more accurately quantified transformation efficiency. The significant difference, calculated from a two-tailed Wilcoxon rank sum test, is indicated by the bracket at the top of the panel. (C) Violin plots showing gene expression, as quantified by qRT-PCR, in the genotypes assayed in panel B. IONPJBJN_00507 is a coding sequence within the PRCI that is absent from RMV7_wt PRCI_dnaN+att::Janus. The six points for each gene correspond to three technical replicate assays on each of two biological replicates. The horizontal line on the violin plot shows the median relative abundance for each gene in each genotype. Across all panels, significance is coded as: p < 0.05, *; p < 0.01, **; p < 10^-3, ***; p < 10^-4, ****. All p values were subject to a Holm-Bonferroni correction within each panel.

To test whether any other part of the PRCI might inhibit transformation, the entire element was removed, either with (RMV7_wt PRCI_dnaN+att::Janus) or without (RMV7_wt PRCI_dnaN::Janus) the flanking att sites. In both cases, a ∼5-fold increase in transformation rates was observed (Fig. 4A). This implied the PRCI encoded an activity that inhibited the activation of the competence system. To identify where this was located within the MGE, four large mutations were generated within the PRCI: one removing the regulatory genes; one removing the regulatory genes and IONPJBJN_00496; one removing the replication genes; and one removing the integration, regulatory and replication genes. However, none of these mutations had such a large effect on transformation rates as the elimination of the entire element (Fig. 4A). This suggested the inhibition of competence induction was not the consequence of a single gene product, but instead the activity of the MGE itself.

A qRT-PCR assay was employed to test whether activation of PRCI_dnaN could affect transformation through upregulating other stress signals within the cell. This found neither ciaR nor manL expression was altered when the PRCI was removed (Fig. 4C). However, transcription of the chaperone regulator hrcA was halved in the absence of the MGE, consistent with the higher hrcA expression detected when the MGE was more active in RMV7_domi pre-CSP (Table S3). This indicated that the repression of competence by PRCI_dnaN could occur indirectly, through MGE-driven elevation of the protein misfolding responses in the cell.

Repression of competence by conserved chaperone proteins

HrcA is encoded by a gene within the dnaK heat shock operon, and it autoregulates through repressing this chaperone-encoding gene cluster through binding the CIRCE motif [75]. In S. pneumoniae, HrcA binding of CIRCE is reduced at elevated temperatures, relieving its repression of the dnaK and groEL operons, enabling a heat shock response [76]. By contrast, Ca²⁺ ions facilitate HrcA-CIRCE motif binding, inhibiting chaperone expression [77]. Both low Ca²⁺ concentrations and extreme temperatures inhibit the induction of competence [78]. To test whether these effects on transformation efficiency were mediated through their reduction in HrcA-CIRCE interactions, the hrcA gene was disrupted with a Janus cassette, then restored, in RMV7_rare.

Disruption of hrcA was confirmed to relieve repression of dnaK and groL expression through qRT-PCR, with no changes in ciaR expression, demonstrating that any effects on transformation were not mediated through CiaRH (Fig. 5A). In the absence of Ca²⁺, the hrcA::Janus mutant had a lower transformation efficiency than either RMV7_rare, or RMV7_rare hrcA restored (Fig. 5B). This difference grew as the Ca²⁺ concentration increased, consistent with HrcA-CIRCE binding increasing transformation efficiency. These effects could be reproduced in S. pneumoniae R6x grown in a chemically-defined medium (Fig. S33). Unfortunately, the essential nature of the chaperones regulated by HrcA prevented any further identification of the specific mechanism by which competence was inhibited.

Figure 5

The regulation of transformation by HrcA in RMV7. (A) Violin plots showing gene expression, as quantified by qRT-PCR, in RMV7_rare and a mutant derivative in which the hrcA chaperone regulator gene was disrupted by a Janus cassette. The six points for each gene correspond to three technical replicate assays of each of two biological replicates. The horizontal line on the violin plot shows the median relative abundance for each gene in each genotype. The absence of any change in ciaR expression demonstrated the effect on transformation was independent of CiaRH. (B) Scatterplot showing the effect of CaCl₂ on transformation efficiency in RMV7_rare and mutants in which hrcA had been disrupted, and then restored. Each point represents an independent transformation assay of one genotype at the indicated CaCl₂ concentration. The best-fitting dose response logistic models are shown, with the shaded areas corresponding to the 95% confidence intervals. (C) Violin plots showing the transformation efficiency of RMV7_wt, RMV7_rare and RMV7_rare hrcA::Janus during normal growth (35 °C) or following a 40 °C heat shock. Each point corresponds to an independent transformation experiment, and the violin plots have a horizontal line indicating the median transformation efficiency of each mutant at each temperature. (D) Combined effects of chaperone and carbon source regulation on transformation efficiency. Results are displayed as in panel C. Wilcoxon rank-sum tests were conducted between all pairs of genotypes. These found both the single mutants, lacking manLMN or hrcA, were significantly less transformable than the parental genotype. Furthermore, the double mutant was less transformable than either single mutant. Across all panels, significance is coded as: p < 0.05, *; p < 0.01, **; p < 10^-3, ***; p < 10^-4, ****. All p values were subject to a Holm-Bonferroni correction within each panel.

Replicating previous observations, a 40 °C heat shock reduced the transformation efficiency of both RMV7_rare and RMV7_wt (Fig. 5C). However, the heat shock increased the transformation efficiency of RMV7_rare hrcA::Janus, demonstrating the competence machinery itself to be resilient to misfolding at elevated temperatures. Hence the effect of elevated temperature on transformation efficiency appears to be mediated through the decreased CIRCE-HrcA interactions. This cannot explain the increased transformation efficiency of the hrcA::Janus cells after heat shock, but the growth curves of these genotypes at 40 °C suggests the loss of hrcA has pleiotropic effects on pneumococcal physiology (Fig. S34).

To check that this regulation was independent of the ManLMN-mediated effects on competence, the transformation efficiency of the RMV7_rare hrcA::Janus manLMN::cat double mutant was compared to that of the progenitor genotype, and the single mutants (Fig. 5D). This demonstrated an approximately five-fold decrease in transformation for each single mutant, and a ∼25-fold reduction for the double mutant. This is consistent with HrcA and ManLMN both activating competence independently.

Discussion

The transformation efficiency of S. pneumoniae is heterogeneous across populations [9,10,79] and over the history of individual strains [8, 80]. While the competence machinery itself is highly conserved across the species [2], studying the regulation of competence in highly-transformable laboratory isolates will inevitably miss mechanisms that affect this process in clinical isolates. S. pneumoniae RMV7 is a representative of GPSC1 [39], which has diversified through homologous recombination exceptionally quickly, enabling vaccine escape and the acquisition of antibiotic resistance loci [7]. However, even within a single isolate of this strain, only a minority of the population were highly transformable in vitro. Phase variable systems enable such phenotypic heterogeneity to persist within a near-isogenic population [27, 32], resulting in a “bet hedging” strategy in RMV7 that mimics the heterogeneity in transformation efficiency of Bacillus subtilis cultures [33,81,82]. Although the competence state is regarded as a stress response analogous to the SOS response in other bacteria [48], this work suggests its activation requires a nutrient-rich milieu and an absence of abiotic stresses, such as elevated temperature. An alternative stress state, observed in RMV7_domi, upregulates the chaperones and proteases needed to cope with protein misfolding, and proteins likely to ameliorate cell wall disruption, without the protein synthesis and energy expenditure of competence induction. A mixture of stress-response strategies may enable the pneumococcus to survive a greater diversity of environmental challenges.

This heterogeneity might be unexpected, given CSP’s role in synchronising competence induction across pneumococcal populations. The competence inhibitor CiaRH inhibits pheromone signalling through induction of five csRNAs, inhibiting intracellular CSP production [53,83,84], and the upregulation of HtrA, which degrades extracellular CSP [54]. If the pheromone were freely diffusing throughout a population, such inhibition of competence would be homogenous across cells. However, the downregulation of manLMN by CiaRH reduces cells’ ability to respond to a CSP stimulus, enabling heterogeneity within a single population.

Despite its major role in regulating the competence machinery, ManLMN was dispensable for efficient growth in rich media. This transporter is highly conserved across S. pneumoniae (Fig. S35; Table S4), and orthologues serve as the main glucose transporter across many Firmicutes, including other streptococci, Lactococcus lactis and Listeria monocytogenes [60]. In S. pneumoniae, it acts as the central metabolic regulator, with the ManL EIIAB component important in controlling the carbon catabolite preferences of S. pneumoniae cells [85]. Loss of ManLMN in Streptococcus mutans was associated with decreased biofilm formation and transformability [86], suggesting its regulatory role is likely to be shared across many streptococci.

While S. pneumoniae encodes a plethora of transporters [26], experiments in chemically-defined media suggest ManLMN may be the only effective means of importing GlcNAc [60]. Crucially, ManLMN’s role as the primary glucose importer means GlcNAc can act as an intracellular signal without being subject to catabolite repression. On importation, GlcNAc is converted to the cytotoxic compound GlcNAc-6-phosphate. This is typically used for cell wall biosynthesis in proliferating cells, but any excess must be converted to fructose-6-phosphate for glycolysis [87]. Hence GlcNAc is a uniquely informative signal of nutrient availability relative to cell proliferation [88]. The increased availability of GlcNAc by V. cholerae has been suggested to signal the colonisation of a copepod’s chitinous surface [89]. Analogously, GlcNAc is a major component of mucins covering the upper respiratory tract [90], although S. pneumoniae is an obligate commensal of the human nasopharynx, and likely derives little ecological information from GlcNAc’s availability from host structures. An alternative interpretation is locally-elevated concentrations of GlcNAc may be indicative of lysis of bacterial cell walls, as suggested in streptomycetes [91]. As V. cholerae and S. pneumoniae both co-ordinate fratricide with their competence machinery [92], this would represent a parsimonious explanation for both using GlcNAc as a signal inducing competence.

These two pathogens also both regulate competence through quorum sensing (Fig. 6), and when combined with elevated GlcNAc concentrations, the paired signals would imply high local densities of conspecific bacteria undergoing cell wall lysis. This would indicate the availability of genomic DNA suitable for integration through homologous recombination. There are further similarities in these species’ GlcNAc-signalling pathways. One shared component is a TfoX protein, present in many bacterial phyla (Fig. S36). In gram-negative bacteria, TfoX affects transcription [72] and responds to 3’,5’-cAMP. It is unclear whether the shorter orthologue found in gram-positive bacteria could fulfil either of these roles, particularly given the apparent absence of 3,5’-cAMP in pneumococci, to whom it is likely toxic [70]. The component of the pneumococcal regulatory system most similar to enzymes generating 3’,5’-cAMP was YjbK_Spn, which belongs to a recently-defined subset of CYTH proteins [93, 94]. However, like its orthologues in B. subtilis [95] and Staphylococcus aureus [70], the function of YjbK_Spn is unknown. Both TfoX and YjbK are conserved in S. pneumoniae (Fig S35), although tfoX is absent from ∼12% of isolates [96]. However, neither is strongly upregulated by CSP, making it difficult to identify their regulatory functions through gene expression studies alone (Fig. S37).

Figure 6

Comparison of the regulation of competence in S. pneumoniae (left), from this work, and V. cholerae (right), summarized from [72]. In each, competence is regulated by a quorum-sensing system: CSP in S. pneumoniae, and the cholera autoinducer 1 (CAI-1) in V. cholerae. CAI-1 operates through inhibiting LuxO, thereby activating the HapR protein, which indicates a high-cell density environment. HapR activates competence through the Quorum-Sensing and TfoX-dependent Regulator (QstR). This regulator also senses the activation of TfoX in response to GlcNAc being sensed by the transmembrane regulator TfoS, via the TfoR small RNA. TfoX activity is also regulated by the Catabolite Regulatory Protein (CRP), which is activated by 3’,5’-cAMP, generated by the CyaA adenylate cyclase under carbon source starvation conditions. Hence there are parallels with the TfoX orthologue, and adenylate cyclase-like protein YjbK, responding to GlcNAc in S. pneumoniae RMV7. GlcNAc also appears to promote competence through a TfoX/YjbK-independent route, based on the behaviour of RMV7_wt. There are no known parallels in V. cholerae for the regulation of competence by HrcA, the binding of which to CIRCE sequences is promoted by Ca²⁺, thereby activating competence through an unknown mechanism. Note in S. pneumoniae, these signals are shown to converge on the competence-specific sigma factor ComX for simplicity, but the mechanisms and targets of the signalling pathways are unknown.

In vitro transformation of S. pneumoniae was also activated by the addition of CaCl₂, with cells most sensitive to changes in Ca²⁺ concentration in the 0-5 mM range. This spans the relevant concentrations of the ion in mucus [97] and interstitial fluid, which rise on cell necrosis and inflammation [98]. Ca²⁺ was previously suggested to aid the translocation of DNA molecules across the plasma membrane [99]. We found its effect was dependent upon HrcA, the binding of which to CIRCE motifs is promoted by higher Ca²⁺ concentrations [77]. Both the deletion of hrcA, and the increased expression in RMV7_domi associated with PRCI-associated stress, were associated with reduced transformation efficiency. This suggests the activation of competence is mediated by HrcA-CIRCE complexes. Any involvement of the chaperones regulated by HrcA could not be tested in this genetic study. In other species, orthologues of the chaperones regulated by HrcA-CIRCE have been found to block sigma factor switching, as required for the activation of late competence genes [100]. However, the change in the expression of chaperone genes in response to alterations in PRCI replication was small compared to that of hrcA. Hence HrcA may represent a physiologically-relevant competence regulator that acts through a chaperone-independent route.

The level of complexity of competence regulation in S. pneumoniae is similar to that in Bacillus subtilis [101], as well as Vibrio cholerae [102] and other gram-negative bacteria [103]. Such signalling networks may be the product of adaptive evolution, selecting for the integration of many external cues to trigger competence only under optimal conditions. However, this would require that S. pneumoniae benefits from transformation when GlcNAc is abundant and there is no protein misfolding stress, whereas H. influenzae benefits during starvation conditions [104–106], when both species are frequently isolated within the same nasopharynx [107]. Perhaps tellingly, the regulatory roles of ManLMN and HrcA are moonlighting functions. These are secondary physiological activities separate from their primary physiological role [108]: the transporter activated transformation in media conditions under which it did not affect growth (Fig. 3), and the chaperone regulator activated transformation without being necessary for the correct folding of the proteins involved (Fig. 5).

Alternatively, the complexity of competence regulation may be a product of the conflict between organism-level adaptation and gene-level selection. The conservation of the competence machinery suggests individual cells must benefit from transformation. Yet any gene in which variation can accumulate is expected to evolve into forms that selfishly reduce transformation efficiency, as this limits the rate at which such loci can be replaced within a recombination recipient by competing alleles from donors [17, 105]. Previously documented for MGEs [6,15–17], this study suggests non-mobile variable loci (e.g. the tvr locus) may also evolve to suppress transformation. Counteracting this effect in pneumococci are the moonlighting conserved regulators ManLMN and HrcA. Such proteins are well-suited to synchronising and activating competence, because purifying selection on their primary function impedes the emergence of selfish mutants that inhibit competence. Such conflicting gene-level pressures may explain why so many loci regulate the competence machinery, and why transformation efficiency is so variable across and within species [109].

Methods and materials

Cell culture and mutagenesis

Genotypes used in this study are described in Table S5. Unless otherwise stated, encapsulated S. pneumoniae were cultured statically at 35 °C supplemented with 5% CO₂ in 10 mL of a mixed liquid media, consisting of a 2:3 ratio of Todd-Hewitt media with 0.5% yeast extract (Sigma-Aldrich), and Brain-Heart Infusion media (Sigma-Aldrich). Transformation experiments with S. pneumoniae R6x derivatives used a chemically-defined medium, consisting of disodium β-glycerophosphate (20 g L^-1; Sigma-Aldrich), sodium pyruvate (0.1 g L^-1; Fluorochem), choline (0.001 g L^-1; Alfa Aesar), cysteine (0.4 g L⁻¹; Tokyo Chemical Industry UK), glucose (3.8 mM; Sigma-Aldrich) and galactose (12 mM; Sigma-Aldrich). Carbohydrates were otherwise added as a supplement to liquid media at a final concentration of 33 mM, unless otherwise specified.

Culturing on solid media used Todd-Hewitt media supplemented with 0.5% yeast extract. For the generation of mutants, 0.8-1 kb PCR fragments of regions flanking the gene of interest were amplified with an added BamHI or EcoRV restriction enzyme site. Oligonucleotide sequences are listed in Table S6. PCR products were digested with the appropriate restriction enzymes (Promega) at 35 °C for 2-4 hours, and then ligated to the chosen antibiotic markers using T4 DNA ligase (Invitrogen). Ligations were used as templates for further amplifications using PCR, which generated the constructs used for mutagenesis through transformation of competent pneumococcal cells.

Transformation assays

One millilitre of bacterial culture was collected at an OD₆₀₀ between 0.15-0.25. Cells were then incubated for 2 hours at 35 °C with 5 μl of 500 mM CaCl₂ (Sigma), 250 ng of competence stimulating peptide 1 (CSP-1; Cambridge Bioscience Ltd) and 100 ng of the purified rpoB gene, containing a base substitution that conferred resistance to rifampicin [110]. After two hours of incubation at 35 °C, a volume of between 1 and 100 μl of the transformed culture was spread on an agar plate supplemented with 4 μg mL^-1 of rifampicin (Fisher Scientific). For precise quantification of transformation frequencies, 1 μl of a 10³-fold dilution of the same culture was spread on a non-selective plate in parallel. Colonies were counted after 24 hours of incubation at 35 °C with 5% CO₂. For other transformations, the same conditions were used, and cells were incubated on agar plates supplemented with the appropriate selective antibiotics: kanamycin (Sigma-Aldrich) at 600 μg mL^-1, or chloramphenicol (Sigma-Aldrich) at 4 μg mL^-1.

A logistic curve was used to analyse the relationship between transformation efficiency (in transformants per 10⁴ cfu), t, and the volume of 25 mM CaCl₂ added to the 1 mL of cells, v. The fitted function was:

The values of the variables a, b and c were estimated using the Levenberg-Marquardt nonlinear least-squares algorithm in the minpack R package [111], using the starting values of 10 (increased to 500 for S. pneumoniae R6x), 1 and 0.5, respectively. The confidence intervals were calculated through refitting the function to 999 bootstrapped samples using the car R package [112].

Surface adhesion assays and growth curves

To measure growth curves, 2×10⁴ cells from titrated frozen stocks were grown in mixed liquid media in 96-well microtiter plates at 35°C with 5% CO₂ for 20 h. Measurements of OD₆₀₀ were taken at 30 minutes intervals over 16 hours using the FLUOstar Omega microplate reader (BMG LABTECH). Three replicate wells were used for each tested genotype in each experiment. For measuring adhesion to abiotic surfaces, at the end of the growth curve incubation, the microtitre plate was submerged in water and dried for 10 minutes. Then 150 μl of a freshly-diluted 0.1% crystal violet solution (Scientific Laboratory Supplies) was added to each well, followed by incubation for 30 minutes at room temperature. Each well was then washed by repeatedly submerging the plate in water to remove excess crystal violet. The plate was incubated at room temperature in an inverted position for four hours. Subsequently, 150 μl of 30% acetic acid (Honeywell) was added to each well. Adherence was quantified as OD₅₅₀ across replicate wells, measured by a FLUOstar Omega plate reader.

Preparations of RNA samples and quantitative PCR

Three replicate cultures of RMV7 tvr_domi::Janus and RMV7 tvr_rare::Janus were grown in 25 mL of mixed liquid media until they reached an OD₆₀₀ of 0.15. A 5 mL sample of bacterial cells was collected and 50 μL 250 ng mL^-1 CSP1 was added to the remaining culture. Further 5 mL samples were taken from each culture 10 and 20 min post-CSP addition. Each sample was immediately treated with 10 mL RNAprotect (Qiagen) and incubated at room temperature for 5 min. Cell were then pelleted by centrifugation at 3,220 g for 10 min. RNA was extracted from the washed pellets using the SV Total RNA Isolation System (Promega) according to the manufacturer’s instructions. The extracted RNA was used for RNA sequencing or qRT-PCR.

All qRT-PCR experiments were conducted as described previously [35], using 0.2 μg of RNA to generate cDNA with the First-Strand III cDNA synthesis kit (Invitrogen). Reactions used the PowerUp™ SYBR™ Green Master Mix (Thermo Fisher) and the QuantStudio™ 7 Flex Real-Time PCR System (Applied Biosystems).

Generation and analysis of RNA-seq data

RNA samples were quantified using an Agilent Bioanalyser RNA Nano Chip (Agilent Technologies), and treated with the RiboZero® rRNA Removal Kit for Bacteria (Illumina) to deplete rRNA. The samples were then cleaned with Agencourt RNAClean Beads (Beckman Coulter). Sequencing libraries were generated with the NEBNext® Ultra II Directional Library Prep Kit for Illumina (New England BioLabs), modified to use oligonucleotide sequences appropriate for the sequencing pipelines of the Wellcome Sanger Institute. The library was amplified through nine PCR cycles using the Kapa HiFi HotStart Ready Mix (Roche) to generate sufficient material for sequencing. All eighteen samples were sequenced as a multiplexed library on a single lane of a HiSeq 4000 sequencing system (Illumina), generating 200 nt paired end reads.

The set of genes used for expression analysis were the 2,088 protein coding sequences annotated on the S. pneumoniae RMV7_domi genome (accession code OV904788), and the 81 non-coding RNAs predicted by infernal version 1.1.2 [113] using the Rfam database [114]. RNA-seq reads were aligned to these sequences using kallisto version 0.46.2 [115], using default settings and 100 bootstraps. Differential gene expression analysis used sleuth version 0.30 [116]. Wald tests were conducted to compare the pre-CSP samples for RMV7_domi and RMV7_rare, and to compare the 10 min and 20 min post-CSP samples for each genotype to the corresponding pre-CSP samples. Visualisation and plotting of data used the genoplotR [117], circlize [118], cowplot [119], ggpubr [120] and tidyverse [121] packages.

Quantification of 3’,5’-cAMP production

Quantification of 3’,5’-cAMP used the Cyclic AMP XP^® assay kit (Cell Signalling Technology). Samples were harvested in the exponential (OD₆₀₀ of 0.2) and stationary (OD₆₀₀ of 0.5) phases of S. pneumoniae statically-grown cultures. E. coli DH5α was grown in 25 mL Luria-Bertani media (Sigma-Aldrich) at 37 °C, shaken at 250 revolutions per minute, and samples were harvested in the exponential (OD₆₀₀ of 0.3) and stationary (OD₆₀₀ of 1.0) phases. Cells were pelleted through centrifugation at 3,220 g for 10 min. The cell pellets were re-suspended in 850 μL of the kit’s lysis buffer and 150 μl of lysozyme (10 mg mL^-1), followed by incubation at 35 °C for 20 minutes. Cells were pelleted by centrifugation at 9,500 g for 5 minutes. The supernatants were collected, and the cell pellets were resuspended in 400 μl phosphate-buffered saline. The protein concentrations of all the collected samples were adjusted to 400 μg mL^-1 using the Qubit protein broad range assay kits (Qiagen). The Cyclic AMP XP^® kit ELISA plates were prepared according to the manufacturer’s protocols, and a 50 μl sample of supernatant from each tested culture was loaded in each well. The OD₄₅₀ was measured using the FLUOstar Omega plate reader. Concentrations of 3’,5’-cAMP were calculated using a standard curve, according to the manufacturer’s instructions.

Analysis of motif distribution and sequence diversity

The overall distribution of methylation motifs was calculated using DistAMo [122]. The calculation of distances between coding sequences and motifs used biopython.

The diversity of genes encoding regulatory proteins within S. pneumoniae used previously-calculated pairwise genetic distances [123]. Only genes found in at least half the isolates in the collection were used, to avoid including genes for which diversity might be inferred to be low, due to a limited samples size for detecting polymorphisms.

Proteins containing the TfoX N-terminal domain were identified using EMBL SMART [124]. These amino acid sequences were aligned with MAFFT [125], and a phylogeny generated with Fasttree2 [126] using default settings. Proteins were assigned to Families using the NCBI Taxonomy [127].

Footnotes

• Added detail in the Methods section.