ABSTRACT
Cyclic di-AMP (c-di-AMP) is a newly discovered bacterial second messenger. However, regulation of c-di-AMP homeostasis is poorly understood. In Streptococcus pneumoniae, a sole diadenylate cyclase, CdaA, produces c-di-AMP and two phosphodiesterases, Pde1 and Pde2, cleave the signaling dinucleotide. To expand our knowledge of the pneumococcal c-di-AMP signaling network, we performed whole-genome sequencing of Δpde1 Δpde2 heat shock suppressors. In addition to their effects on surviving heat shock, these suppressor mutations restored general stress resistance and improved growth in rich medium. Mutations in CdaA or in the potassium transporter TrkH paired with an insertion leading to a frameshift at the C terminus of CdaA significantly reduced c-di-AMP levels. These observations indicate that the elevated c-di-AMP levels in the Δpde1 Δpde2 mutant enhance susceptibility of S. pneumoniae to the stress conditions. Interestingly, we have previously shown that TrkH complexes with a Trk family c-di-AMP-binding protein, CabP, to mediate potassium uptake. In this study, we found that deletion of cabP significantly reduced pneumococcal c-di-AMP levels. This is the first observation that a c-di-AMP effector protein modulates bacterial c-di-AMP homeostasis.
IMPORTANCE Second messengers, including c-di-AMP, are prevalent among bacterial species. In S. pneumoniae, c-di-AMP phosphodiesterase-encoding gene null mutants are attenuated during mouse models of infection, but the role of c-di-AMP signaling in pneumococcal pathogenesis is enigmatic. In this work, we found that heat shock suppressor mutations converge on undermining c-di-AMP toxicity by changing intracellular c-di-AMP concentrations. These mutations improve the growth and restore the stress response generally in c-di-AMP phosphodiesterase-deficient pneumococci, thereby demonstrating the essentiality for tight regulation of c-di-AMP homeostasis in order to respond to stress. Likewise, this work demonstrates that a c-di-AMP effector protein, CabP, affects c-di-AMP homeostasis, which provides new perception into c-di-AMP regulation. This study has implications for c-di-AMP-producing bacteria since many species contain CabP homologs.
INTRODUCTION
Streptococcus pneumoniae (the pneumococcus) is a commensal bacterial species in the nasopharynx, but is pathogenic under some circumstances and can cause pneumonia, bacteremia, sinusitis, otitis media, and meningitis (1–4). As such, this deadly pathogen needs to be able to survive under diverse stress conditions, including changes in pH (from mildly acidic at inflammatory foci to alkaline during dissemination) and osmotic, oxidative, and heat stresses (5–7). Regarding heat stress, as the bacteria invade tissues, environmental temperature can change from 30 to 34°C to 37°C or greater during fever (8). Since pneumococci are adept at enduring environmental stresses and have increased incidences of antibiotic resistance and vaccine evasion (9–14), a better understanding of pneumococcal physiology is needed to combat pneumococcal disease.
Cyclic di-AMP (c-di-AMP) is a recently discovered signaling molecule. Diadenylate cyclase enzymes convert two ATP molecules to c-di-AMP (15, 16), while c-di-AMP phosphodiesterases break down c-di-AMP to either phosphoadenylyl adenine (pApA) or AMP (17, 18). The diadenylate cyclase domain is widespread among nearly 1,500 archaeal and bacterial species, but is absent in eukaryotic genomes (19), which allows the c-di-AMP signaling network to be an attractive target for therapeutics. Diadenylate cyclase-encoding genes are essential in many bacteria, including in S. pneumoniae (20, 21). In Bacillus subtilis, which contains three diadenylate cyclase genes, each diadenylate cyclase gene could be individually deleted, but all three could not without complementation (22). In a more recent study, deletion of all three diadenylate cyclase genes was rescued either by growing the mutant in a medium with low [K+] or by a suppressor mutation in NhaK, a cation exporter (23). The expression of B. subtilis CdaA, a homolog of pneumococcal CdaA (previously named DacA, which conflicts with the name for d-alanyl–d-alanine carboxypeptidase), is downregulated under low-[K+] conditions, consistent with the relationship between c-di-AMP production and potassium requirements. Additionally, the sole diadenylate cyclase in the pathogen Listeria monocytogenes could be removed when grown in minimal medium or concurrent with suppressor mutations in relA, which synthesizes the stringent response alarmone (p)ppGpp (24). In Staphylococcus aureus, not only do high c-di-AMP levels correlate with an increase in (p)ppGpp levels, but the c-di-AMP phosphodiesterase GdpP is inhibited by ppGpp (25). These findings suggest that c-di-AMP levels are modulated through environmental inputs and are associated with the second messenger (p)ppGpp.
Bacterial second messengers such as c-di-AMP relay signals to effector molecules after sensing environmental changes. Within bacteria, c-di-AMP has been shown to affect peptidoglycan homeostasis (22, 26–29), transcription through the ydaO family of riboswitches (30–32), osmolyte transport (33, 34), and potassium transport (35–37). Specifically, KtrA and CpaA of S. aureus and CabP of S. pneumoniae are all RCK (regulator of conductance of K+) domain proteins that bind c-di-AMP and control potassium uptake (35, 36, 38, 39). c-di-AMP is also recognized by the histidine sensor kinase KdpD, which initiates the expression of the potassium transport Kdp system in S. aureus (35, 37). In B. subtilis, c-di-AMP controls expression of the potassium transporter KimA via the ydaO riboswitch (23, 30–32). Additional target proteins have been characterized in recent years. The pyruvate carboxylase in L. monocytogenes (40) and Lactococcus lactis (41) is inhibited by c-di-AMP, which modulates aspartate and citrate metabolism (41, 42). The osmolyte transporter OpuC was found in L. monocytogenes (33) and S. aureus (34) to be impeded by c-di-AMP binding, disrupting carnitine solute import. In low [K+] without the main high-affinity transporters that are c-di-AMP responsive, cell viability was restored by increases in positively charged arginine biosynthesis intermediates (43). Together with controlling potassium uptake, c-di-AMP plays a concerted role in regulating osmotic pressure (44, 45).
Consistent with the need for intricate control of c-di-AMP homeostasis, fluctuations of c-di-AMP levels can mediate survival, virulence, and the stress response (46). Related to osmolyte and ion transport, osmotic force susceptibility has been reported with increased c-di-AMP levels and resistance in organisms in which the diadenylate cyclase-encoding genes were mutated (28, 33, 34, 42, 47–49). Curiously, deletion of c-di-AMP phosphodiesterase gdpP in L. lactis enhances heat shock resistance and makes the organism more sensitive to osmotic stress (48). Dysregulation of c-di-AMP signaling can affect tolerance to antibiotic agents. β-Lactam resistance has been shown to correlate with c-di-AMP levels (26, 27, 42, 50–54), suggesting a role for c-di-AMP in cell wall metabolism. With reference to DNA damage, the activity of the diadenylate cyclase DisA is inhibited upon recognition of branched DNA in B. subtilis (15), which can delay sporulation, but an increase in c-di-AMP levels can allow cells to continue sporulating (55). Likewise, more studies have continued uncovering a mechanism for c-di-AMP signaling to aid in the DNA damage response from oxidative stress or other DNA-damaging agents (17, 56–58).
S. pneumoniae contains genes that encode one diadenylate cyclase, CdaA, and two c-di-AMP phosphodiesterases that hydrolyze this dinucleotide to either pApA (Pde1) or AMP (Pde2) (17). Both phosphodiesterases contain DHH/DHHA1 domains, which are required for their activity. These c-di-AMP hydrolases are each indispensable for pneumococcal virulence (17, 59). Strains containing deletions of one or both phosphodiesterases have elevated intracellular levels of c-di-AMP. Previously, we noted that these strains have a growth defect and are more susceptible to UV DNA damage (17). In regard to c-di-AMP effector protein function in S. pneumoniae, the c-di-AMP-binding protein CabP interacts with potassium transporter TrkH (SPD_0076) to facilitate potassium uptake. However, c-di-AMP impairs this process upon binding with CabP (36).
In this study, we determined that c-di-AMP regulation is essential to survive acidic, osmotic, and heat stresses in S. pneumoniae D39. In particular, we uncovered mutations that specifically link pneumococcal c-di-AMP homeostasis to the stress response. We demonstrated that either mutation of a critical residue in CdaA or addition of a short tag at the C terminus of CdaA lowered its enzymatic activity, which increased stress resistance in the Δpde1 Δpde2 mutant. Moreover, we found that the c-di-AMP-binding protein CabP modulates c-di-AMP homeostasis.
RESULTS
c-di-AMP phosphodiesterase mutants are highly susceptible to heat shock, acidic, and osmotic stresses.We previously reported that the proteins encoded by pde1 and pde2 are c-di-AMP phosphodiesterases that hydrolyze c-di-AMP to pApA and AMP, respectively. Strains with deletions of these genes contain elevated intracellular c-di-AMP concentrations. We found that these strains were more sensitive to UV treatment than the wild type (WT) and had a significant growth defect in Todd-Hewitt broth containing 0.5% yeast extract (THY broth) (17). In the present study, we explored whether these mutants are susceptible to other stress environments, including heat shock, acidic, and highly osmotic conditions. We incubated the WT and three phosphodiesterase mutants (Δpde1, Δpde2, and Δpde1 Δpde2 mutants) at 50°C for 30 min and then spotted aliquots of serially diluted bacteria onto plates containing tryptic soy agar plus 5% sheep blood (TSAB) or directly spotted untreated dilutions onto TSAB plates adjusted to pH 6.0 either with HCl or with the addition of 0.25 M NaCl. We also spotted untreated samples onto regular TSAB plates as controls. We recovered significantly fewer CFU of the Δpde1, Δpde2, and Δpde1 Δpde2 mutants compared to the WT on all of the stress-conditioned plates but similar numbers of each bacterial strain on the control plates (Fig. 1; see Fig. S1 in the supplemental material), suggesting that both Pde1 and Pde2 are essential for overcoming environmental stressors. As Pde1 and Pde2 are distinct proteins and their catalytic products of c-di-AMP differ, it is very likely that accumulation of high levels of c-di-AMP is toxic and prevents adaptation to these stress conditions in S. pneumoniae.
Susceptibility of c-di-AMP phosphodiesterase mutants to heat shock and acid stress. (A) Equivalent numbers of cells of the S. pneumoniae WT and pde mutants were serially diluted 10-fold in THY medium, and 10 μl was spotted onto TSAB plates as a control. Dilutions are indicated. For heat shock, dilutions were incubated for 30 min for 50°C prior to spotting onto agar plates. For acidic stress, bacteria were spotted onto TSAB plates adjusted to pH 6.0 with HCl. All plates were incubated overnight prior to imaging. (B and C) Survival rates after heat shock (B) and acid stress (C) were calculated. The images shown in panel A are representative of three independent replicate experiments. Data shown in panels B and C are the means from three independent replicate experiments. The error bars indicate the standard errors of the means (SEMs). **, P < 0.005; ***, P < 0.001; ****, P < 0.0001.
Three classes of suppressors are recovered from Δpde1 Δpde2 bacteria after heat shock.Since there were very few Δpde1 Δpde2 colonies present after heat shock treatment, it is likely that Δpde1 Δpde2 survivors were selected for enhanced resistance to heat stress. Therefore, we anticipated that sequencing of heat shock suppressor mutants would provide novel information in the pneumococcal c-di-AMP signaling network. We screened two rounds for heat shock-resistant (HSR) colonies. A screening of 8 colonies in the first round and 24 colonies in the second round yielded 4 and 12 HSR strains, respectively (∼50% are suppressors) (Fig. 2). Whole-genome sequencing of nine of these stable suppressor strains compared to the Δpde1 Δpde2 mutant yielded three specific classes of mutations (Table 1): a V76G mutation in CdaA (CdaA*), a C-terminal truncation in PhoU2 (PhoU2Δ40), and an S69A mutation in TrkH (TrkH*) paired with an insertion of an additional adenine at the 3′ end of cdaA, which leads to a frameshift and a C-terminal tail of 32 amino acids (aa) (CdaA-32) (see Fig. S3 in the supplemental material for amino acid sequence and CdaA protein expression). Since the mutations within each class are identical, it is likely that the mutants of each class are siblings.
Screening of heat shock suppressors in the Δpde1 Δpde2 mutant. After an initial heat shock experiment, Δpde1 Δpde2 survivors (HSRs) were grown to make new stocks. The same numbers of bacteria were serially diluted in THY and incubated for 30 min at 50°C to assess their heat shock resistance. Dilutions are indicated below the graphs. Both untreated and heat-treated dilutions were spotted onto TSAB plates and incubated overnight. A total of 32 candidates were evaluated. Representative images of the first eight clones are shown. The survivors after initial screening were further characterized for response to heat shock and acid stress conditions. We noticed that HSR1, -2, -5, and -8 exhibited similar outcomes to heat shock and low-pH conditions (see Fig. S2 in the supplemental material).
Heat shock suppressor screening
First, we wanted to test if these mutations specifically increased stress resistance in the Δpde1 Δpde2 mutant. We took one representative HSR isolate for each category and subjected the strains to heat shock, as well as acidic media, and compared the survival of these isolates to the WT and Δpde1 Δpde2 strains. All three classes of mutations increased stress resistance in general. Not only did the suppressor strains subsist better after heat shock, they had higher survival rates under acidic conditions than the Δpde1 Δpde2 strain (Fig. 3A). Previously, we found that the Δpde1 Δpde2 mutant had a growth defect compared to the WT in THY broth (17). Since the suppressor mutants decreased stress susceptibility in general, we expected that they may improve the growth rate as well. We monitored the turbidity of the suppressor mutants grown in THY broth. All of the suppressor strains showed partially restored growth to WT levels (Fig. 3B). The WT strain had a doubling time of 49.6 ± 2.4 min, while the Δpde1 Δpde2 strain had a significantly higher doubling time of 89.1 ± 10.3 min. The suppressor strains grew with mean doubling times of 57.2 ± 11.1, 66.4 ± 9.1, and 63.4 ± 5.0 min for HSR1, -2, and -16, respectively, which were slightly lower than that of the Δpde1 Δpde2 strain. Taking these results together, we found that while we recovered three classes of suppressor mutations, they all converge on improving the stress response and growth of the Δpde1 Δpde2 mutant.
Characterization of stress resistance, growth, and c-di-AMP levels of the selected HSR strains. (A) The three representative HSR strains from each group (Table 1) were serially diluted in THY, and 10 μl was spotted onto TSAB control plates or plates adjusted to pH 6.0 and incubated overnight. Heat shock dilutions were incubated for 30 min at 45°C prior to spotting onto control agar. The data shown are representative of three independent biological replicate experiments. (B) Bacterial growth of HSR mutants. The S. pneumoniae WT, Δpde1 Δpde2 mutant, and representatives of each HSR mutation group were inoculated into THY broth and monitored at OD620. The HSR strains all followed the same growth pattern and are indistinguishable. The error bars denote the standard deviations (SD) from three independent experiments. (C and D) Pneumococcal strains as indicated were grown to an OD620 of 0.4. Lysates of bacterial pellets were assayed for intracellular c-di-AMP levels (C), and culture supernatants were collected to determine secreted c-di-AMP (D) by ELISA. (E) Recombinant E. coli strains harboring the vector control, a plasmid expressing either WT pneumococcal CdaA, CdaAV76G, or CdaA-32 were grown overnight, and bacterial lysates were assayed for intracellular c-di-AMP. Data shown in panels C to E are the means from three independent replicate experiments. The error bars indicate the SEMs. **, P < 0.005; ***, P < 0.001; ****, P < 0.0001.
Heat shock suppressor mutations alter c-di-AMP levels compared to the Δpde1 Δpde2 strain.It is known that the Δpde1 Δpde2 mutant contains elevated levels of intracellular c-di-AMP (17) (Fig. 3C), so we anticipated through the heat shock screening to obtain mutations involved in either c-di-AMP homeostasis that decrease c-di-AMP levels or a c-di-AMP effector protein that alleviates the sensitivity to heat shock. The screening revealed two distinct mutations in the sole diadenylate cyclase CdaA, resulting in either CdaA* or CdaA-32. We hypothesized that these mutations may lower c-di-AMP production in order to survive the stress conditions. However, the other class also restored general stress resistance and growth similarly to those of the Δpde1 Δpde2 cdaA* strain, so we measured intracellular and extracellular c-di-AMP levels for all three suppressor mutation classes and compared them to the WT and the Δpde1 Δpde2 strain. In L. monocytogenes, several multidrug efflux pumps are responsible for secreting c-di-AMP (60). However, in S. pneumoniae the c-di-AMP secretion mechanism has not been elucidated, but extracellular c-di-AMP can be detected (Fig. 3C). As expected, cdaA* significantly reduced c-di-AMP levels (∼8-fold reduction) (Fig. 3C). Streptococcus pneumoniae phoU2 encodes a phosphate signal transduction protein (61). Remarkably, phoU2Δ40 slightly increased c-di-AMP levels (∼1.7-fold), while trkH* paired with cdaA-32 significantly decreased c-di-AMP (∼8-fold) in the Δpde1 Δpde2 background (Fig. 3C). We found that the comparative levels of intracellular c-di-AMP matched those of the secreted samples (Fig. 3D), suggesting that these strains do not have altered c-di-AMP secretion but rather have altered production of c-di-AMP.
In order to further examine whether the V76G and 32-aa-addition variations alter the enzymatic activity of CdaA, we expressed CdaA, CdaAV76G, and CdaA-32 individually in Escherichia coli and determined c-di-AMP production in the recombinant E. coli strains. It is well known that E. coli does not produce c-di-AMP (62), and therefore the c-di-AMP levels we examined reflect the activity of CdaA. We found that both the V76G substitution and 32-aa extension significantly decreased the enzymatic activity of CdaA (Fig. 3E), which supports that the mutations in CdaA account for the majority of the reduced levels of c-di-AMP determined from the HSR strains. These results also indicate that the elevated c-di-AMP levels in the Δpde1 Δpde2 strain enhance susceptibility of S. pneumoniae to the stress conditions. Additionally, the phoU2 suppressor mutation provided better fitness for the bacteria under the tested conditions with an unknown mechanism, even though these bacteria possess high levels of c-di-AMP.
The effect of TrkH point mutation on c-di-AMP homeostasis in the suppressor mutants.TrkH is a potassium transporter that complexes with the c-di-AMP effector protein CabP in order to take up potassium; however, potassium transport is impaired when CabP binds c-di-AMP and dissociates from TrkH (36). In a recent study, the expression of CdaA in B. subtilis was shown to be influenced by environmental potassium conditions (23). We hypothesized that the TrkH* mutation recovered from the suppressor screening may also afford a benefit by influencing potassium uptake and/or by altering c-di-AMP homeostasis. To test whether TrkH* plays a role in response to the stress conditions, we generated mutations involving trkH*, cdaA-32, and both variations, respectively, in the Δpde1 Δpde2 background. We compared the stress response and growth of these bacterial strains to those of the WT and Δpde1 Δpde2 strains and the isolated suppressor mutant HSR16. After heat shock, the strains containing the cdaA-32 mutation in the Δpde1 Δpde2 background survived just as well as WT bacteria (Fig. 4A). Similarly, these strains survived better than the Δpde1 Δpde2 strain under acidic conditions, while trkH* alone in the Δpde1 Δpde2 strain had a minor effect on stress resistance (Fig. 4A). However, trkH* had a similar effect to cdaA-32, and both mutations partially restored bacterial growth (Fig. 4B). In order to determine whether TrkH* affects bacterial potassium uptake, we generated trkH* and trkH* ΔcabP mutations in the WT and grew these strains in a chemically defined medium (CDM) with various concentrations of potassium. The results showed that the trkH* strain grew similarly to the WT under either high- or low-K+ conditions, but the trkH* ΔcabP strain did not grow in media with low K+ concentrations (see Fig. S4 in the supplemental material), which suggests that TrkH* still can complex with CabP and requires CabP to take up potassium.
Stress resistance and c-di-AMP levels in strains possessing CdaA-32 and TrkHS69A variations. (A) Dilutions of the indicated strains were plated on TSAB agar plates (control) or TSAB agar plates adjusted to pH 6.0. Heat shock dilutions were incubated for 30 min at 45°C prior to spotting onto TSAB agar. Plates were incubated overnight prior to imaging. The data shown are representative of three independent experiments. (B) Bacterial growth of cdaA-32 and trkH* mutants. The indicated strains were inoculated in THY broth and monitored hourly at OD620. The strains containing cdaA-32 and/or trkH* mutations all followed the same growth pattern and are indistinguishable. The error bars denote the standard deviations (SD) from three independent experiments. (C) Determination of the intracellular c-di-AMP levels of the indicated strains. The data shown are the means of three independent replicate experiments. The error bars indicate the SEMs. ****, P < 0.0001.
Next, we decided to investigate the role of these mutations in c-di-AMP homeostasis. We measured intracellular c-di-AMP in the same strains that were probed for the stress response. While the cdaA-32 mutation accounted for the majority of the reduction in c-di-AMP in HSR16, the addition of trkH* alone in the Δpde1 Δpde2 mutant also slightly lowered intracellular c-di-AMP levels (30% reduction) (Fig. 4C). Furthermore, the introduction of trkH* to the Δpde1 Δpde2 cdaA-32 strain additively reduced c-di-AMP levels from 22.8 ± 0.8 pmol/OD620 (optical density at 620 nm) to 16.0 ± 0.7 pmol/OD620. In all strains tested, there was a clear correlation between intracellular c-di-AMP and survival after heat shock. Taking together results for stress survival and effect on c-di-AMP amounts, we conclude that in HSR16, cdaA-32 plays a major role, while trkH* has a minor contribution to reducing c-di-AMP levels and improving stress resistance.
CabP controls pneumococcal c-di-AMP homeostasis.Since trkH* affected c-di-AMP homeostasis in the Δpde1 Δpde2 strain, we hypothesized that TrkH and CabP may control pneumococcal c-di-AMP homeostasis. We measured c-di-AMP in the WT and trkH and/or cabP mutants, and the trkH* strain. Deletion of cabP markedly reduced c-di-AMP compared to the WT strain (Fig. 5A). The decrease in c-di-AMP levels in the ΔcabP mutant was restored by expressing cabP under the promoter of the trkH-cabP operon using a shuttle plasmid (Fig. 5B), indicating the specificity of CabP in controlling pneumococcal c-di-AMP homeostasis. The reduction in c-di-AMP by deletion of cabP was also detected from all the pde mutants (see Fig. S5 in the supplemental material), which rules out Pde1 and Pde2 being affected by CabP to alter c-di-AMP levels. Surprisingly, ΔtrkH and trkH* did not decrease c-di-AMP levels (Fig. 5A and C), indicating that trkH* only alters c-di-AMP levels when bacteria possess high levels of this dinucleotide. Given that TrkH and CabP are both required for potassium uptake (36), our result indicates that CabP controls c-di-AMP levels in a potassium uptake-independent manner.
Deletion of cabP and trkH and their effect on c-di-AMP levels. Intracellular c-di-AMP samples were prepared and detected by ELISA. (A) Determination of c-di-AMP levels of the strains indicated. (B) Complementation of cabP under its native promoter with the shuttle plasmid pVA838. (C) Comparison of c-di-AMP levels between the WT and trkH* mutant. The data shown in all panels are the means from three independent replicate experiments. The error bars indicate the SEMs. *, P < 0.05; **, P < 0.005; ****, P < 0.0001.
In B. subtilis, the reduction of CdaA protein in media with low potassium concentrations was reported to coincide with lower c-di-AMP production (23). This interesting phenomenon was investigated in S. pneumoniae. We found that pneumococcal CdaA expression was also affected by the environmental potassium concentration (see Fig. S6 in the supplemental material). To investigate if CabP affects c-di-AMP homeostasis through differential expression of CdaA, we analyzed the cdaA promoter activity using a promoter-reporter fusion assay and examined CdaA protein levels in the WT, ΔcabP, and ΔtrkH strains by Western blotting. To evaluate the expression of cdaA at the transcriptional level, the promoter of cdaA was cloned into a promoterless lacZ transcriptional fusion reporter plasmid. Empty plasmid was utilized as a negative control, whereas gyrA promoter was utilized as a constitutively expressing control for the assay. Analysis of the β-galactosidase activity of both gyrA and cdaA promoters yielded no change in the mutant strains compared to the WT (see Fig. S7A in the supplemental material). In the detection of CdaA protein via Western blotting, protein expression was consistent among the WT, ΔcabP, ΔtrkH, and ΔtrkH ΔcabP strains (Fig. S7B). Collectively, these results demonstrate that in the ΔcabP strain, the reduced c-di-AMP level is not due to alteration of CdaA expression. Thus, we proposed the possibility of posttranslational control of CdaA by CabP. We hypothesized that CabP may directly interact with CdaA to control c-di-AMP production. To test this hypothesis, we performed a bacterial two-hybrid assay. If the two fusion proteins bind, an active adenylate cyclase will form, and as a result there will be a cAMP-responsive induction of β-galactosidase. As evident in the leucine zipper-positive control, upon cleavage of X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) substrate, the colonies will be blue (see Fig. S8A in the supplemental material). The strain expressing both CdaA and CabP did not show indication of a direct protein-protein interaction, similar to the negative controls (Fig. S8A). However, the bacterial two-hybrid assay may impede binding through the fusion of proteins to the adenylate cyclase domains. As an alternative approach, we coexpressed cdaA and/or cabP in E. coli to determine if CabP can alter the activity of CdaA directly. In this method, the open reading frames (ORFs) were expressed without any tags to hinder binding. Since expression of cdaA in E. coli accumulates a large amount of c-di-AMP, the strains were constructed to express pde2 as well so that the recombinant strains could maintain relatively physiological levels of c-di-AMP. As expected, the coexpression of cdaA and pde2 significantly decreased c-di-AMP levels from expression of cdaA alone. The addition of cabP to this strain did not change c-di-AMP production (Fig. S8B). Collectively, these data suggest a posttranslational control of CdaA via the c-di-AMP-binding protein CabP, which is not through a direct interaction between CdaA and CabP.
DISCUSSION
S. pneumoniae encounters a variety of stressors, including fluctuations in nutrient availability, a range of temperatures as it disseminates from the nasopharynx into other tissues (30 to 37°C and higher during fever), and changes in pH in the extracellular environment, osmolarity, and oxidation conditions (5–8, 63). Elegant control of c-di-AMP production and breakdown is necessary to be able to respond efficiently to stressors, since elevated c-di-AMP has been shown to cause growth defects (17, 22, 24, 26, 27, 29, 49, 64), salt sensitivity (28, 33, 44, 48, 49), and β-lactam resistance (26, 27, 48–53, 65, 66). In this study, we establish that Δpde1 Δpde2 pneumococci are highly susceptible to heat shock, acidic stress, and osmotic stress conditions. Utilizing heat shock survival as a selection tool for suppressors, we found three specific mutations reducing bacterial c-di-AMP levels that enhanced stress resistance as well as restored growth compared to Δpde1 Δpde2 bacteria. Two of these mutations were in cdaA, which is the sole diadenylate cyclase gene identified in pneumococci. Previously, truncations that ablate enzyme activity in cdaA encoded by persistors of phosphodiesterase-deficient B. subtilis established that high levels of c-di-AMP are detrimental for growth (67). Together, these observations support our hypothesis that c-di-AMP dysregulation in S. pneumoniae directly affects the physiology and the stress response of this pathogen.
How high levels of c-di-AMP mediate heat shock and acid stress sensitivity is still unknown. S. pneumoniae contains genes that encode several proteins necessary for heat shock survival, such as those found in the CtsR regulon, including ATP-dependent Clp proteases, chaperones DnaK and GroEL (8, 68–70), and a protein named for its high temperature requirement, HtrA (71, 72). These proteins are involved not only in heat stress but also in oxidative stress, competence, and virulence. It will be interesting to elucidate how c-di-AMP interconnects with pneumococcal stress programs or responsive proteins. In a recent study, deletion of a gene encoding a c-di-AMP phosphodiesterase homolog of Pde1 in L. lactis conferred increased heat resistance and susceptibility to osmotic stress (48). The hypersensitivity to osmotic stress in pde-deficient pneumococci and lactococci is consistent with the decreased osmolyte uptake hypothesis proposed by Commichau et al. (44). In contrast to L. lactis, pneumococcal pde-deficient strains are highly sensitive to heat shock. Of note, the high temperatures regarded as stressful are 37.5°C for L. lactis, a cheese producer, and short term at 45°C for S. pneumoniae, a formidable pathogen that grows well at 37°C (48). Therefore, the L. lactis or S. pneumoniae c-di-AMP-responsive proteins involved in heat resistance may diverge or be active at different temperatures. In terms of acid stress conditions, both Δpde1 and Δpde2 strains did not survive as well as the WT (Fig. 1). In S. aureus, both a gdpP (homologous to pde1) mutant and a ybbR mutant were shown to contain slightly elevated c-di-AMP levels, with one having acid resistance and the other exhibiting extreme susceptibility, respectively (64). Therefore, the amount of the global c-di-AMP pool is not the only determinant of stress tolerance.
In addition to the CdaA mutations, we found a point mutation in TrkH that provided a minor contribution to c-di-AMP reduction as well as stress survival in the Δpde1 Δpde2 mutant. However, an alteration of c-di-AMP levels by trkH* in a WT genetic background was not detected, suggesting that TrkH may only affect c-di-AMP levels under certain conditions. Recently, a report from Gundlach et al. determined that the expression and c-di-AMP production of the diadenylate cyclase CdaA were dependent on environmental potassium levels (23). Similarly, pneumococcal CdaA expression is also responsive to potassium levels (Fig. S6). We reported previously that the potassium transporter TrkH complexes with a c-di-AMP-binding protein, CabP, in order to take up potassium (36). This led us to investigate the involvement of these c-di-AMP-responsive proteins in c-di-AMP homeostasis. Surprisingly, deletion of only cabP, but not trkH, significantly lowered c-di-AMP levels in S. pneumoniae (Fig. 5), indicating regulation of c-di-AMP production by CabP in a potassium-independent manner. The mechanism involving CabP is distinct from the altered potassium concentrations affecting CdaA expression in that intracellular c-di-AMP was measured from bacteria grown in rich media where the potassium concentration is not limiting and it does not affect CdaA protein levels (Fig. S7). Due to a similar reduction in c-di-AMP levels by deletion of cabP in the pde-deficient strains (Fig. S5), it is more likely that CabP regulates c-di-AMP production through CdaA. Based on this study, we propose that CabP may affect pneumococcal c-di-AMP homeostasis (i) by directing allosteric binding to CdaA through TrkH as well as an unknown factor or factors and/or (ii) an alternative function that modulates signal input for CdaA (Fig. 6). So far, posttranslational regulation of CdaA homologs has been observed through CdaR and GlmM in other bacteria, which is thought to occur through specific interactions with the cyclase (22, 28, 73). Our results extend the notion that CabP contributes additional physiological roles besides potassium uptake that will further develop our understanding of the pneumococcal c-di-AMP network. Furthermore, it is motivating to elucidate the other functions and binding partners of CabP in S. pneumoniae that may affect c-di-AMP regulation.
Hypothetical model of pneumococcal c-di-AMP homeostasis. The diadenylate cyclase CdaA converts ATP to c-di-AMP, while the phosphodiesterases Pde1 and Pde2 cleave c-di-AMP into pApA and AMP, respectively. The effector protein CabP binds c-di-AMP, which impedes potassium uptake by the transporter TrkH. Either through an additional effector function or by complexation with an unknown protein, CabP affects the activity of CdaA and modulates c-di-AMP homeostasis.
As a result of the screening, we located a cluster of mutations in the gene encoding PhoU2 and the Pst1 system, which is responsible for inorganic phosphate (Pi) uptake (61). Curiously, these mutations improved the stress response but also increased c-di-AMP levels further. These findings indicate that the increased resistance to stress conditions by these mutations is possibly due to obstruction of a related c-di-AMP signaling pathway or an alteration of Pi uptake that directly affects the stress response. Expression of the pneumococcal pst1 operon is controlled by the inhibitory interaction between PhoU2 and the two-component system PhoR (histidine kinase) and PhoB (response regulator) (61). Previously, it was reported in S. pneumoniae that ΔphoU2 significantly increases phosphorylated PhoB levels (indicating activation) as well as an increase in pst1 operon transcript, signifying a derepression of the Pst1 regulon (61). In E. coli with a C-terminal transposon insertion in phoU, there was a loss of function in repressing the Pho regulon responsible for Pi uptake (74). We hypothesized that the phoU2Δ40 suppressor mutation may be increasing Pi uptake in the Δpde1 Δpde2 strain by derepressing the pst1 operon. This was tested by utilizing a Pi uptake assay. The phoU2Δ40 suppressor strain HSR2 significantly increased Pi uptake compared to both the WT and Δpde1 Δpde2 strains (see Fig. S9A in the supplemental material). This result suggests that the C-terminal truncation of PhoU2 increases Pi acquisition, likely through loss of function in inhibiting PhoB activation. Concurrently, the increase in c-di-AMP is likely a by-product of the beneficial aspect of increasing Pi uptake to overcome stress susceptibility. To this end, intracellular ATP levels were measured by a specific luciferase assay in the WT and Δpde1 Δpde2 strains, as well as the three classes of suppressor strains. We found that the Δpde1 Δpde2 strain and the suppressor strains possessed similar intracellular ATP concentrations to the WT. However, as expected, phoU2Δ40 significantly increased ATP levels (∼1.6-fold increase) (Fig. S9B). Interestingly, there is a second pneumococcal Pst system that is encoded upstream of phoU2, named Pst2. The function of Pst2 has been shown to be negatively regulated by PhoU2 (61). To test if ΔphoU2 enhances heat shock resistance specifically, we subjected the WT, ΔphoU2, and Δpst2-phoU2 strains to heat shock treatment. Our result showed that the ΔphoU2 strain survived the stress condition significantly better than the WT or Δpst2-phoU2 strain (Fig. S9C and D), suggesting that the function of the Pst2 transporter is required for enhanced resistance. We conclude that the C-terminal truncation of PhoU2 aids in increasing Pi uptake, accumulating more intracellular c-di-AMP and ATP, and enhances heat shock resistance in a Pst2-dependent manner. We will further determine whether this mutation affects the c-di-AMP signaling network in a future study.
In summary, we determined heat shock suppressor mutations that subvert c-di-AMP homeostasis in phosphodiesterase-deficient pneumococci, which improved the overall stress response. Additionally, we uncovered a novel regulation of c-di-AMP levels by the pneumococcal effector protein CabP.
MATERIALS AND METHODS
Bacterial strains and growth conditions.The bacterial strains utilized in this study are listed in Table S1 in the supplemental material. S. pneumoniae D39 (serotype 2; ATCC) and its derivatives were grown in Todd-Hewitt broth containing 0.5% yeast extract (THY). Frozen bacterial stocks were prepared by growing cells in THY to OD620 of 0.4 and saving them in THY containing 10% glycerol at −80°C. For growth on plates, tryptic soy agar with 5% sheep blood (TSAB) was used. All S. pneumoniae strains were grown at 37°C with 5% CO2. Antibiotics were supplemented at 4 μg·ml−1 erythromycin, 200 μg·ml−1 kanamycin, or 125 μg·ml−1 streptomycin for pneumococcal strains as needed for selection. E. coli strains were streaked overnight on Luria-Bertani (LB) agar plates, and single colonies were inoculated in LB broth shaking at 220 repetitions per minute at 37°C. Antibiotics were supplemented at 100 μg·ml−1 ampicillin or 25 μg·ml−1 kanamycin for E. coli strains as needed.
Heat shock treatment and screening.Frozen pneumococcal stocks were thawed at room temperature and serially diluted to 10−5 in THY medium at room temperature and subject to heat shock at 45 or 50°C as indicated for 30 min prior to spotting 10 μl onto TSAB plates. As controls, dilutions were directly spotted onto plates. For screening of suppressors, the Δpde1 Δpde2 strain was heated for 30 min at 50°C and then spread onto TSAB agar plates. Individual surviving colonies were grown in THY medium to make new frozen stocks. The heat shock-resistant (HSR) variants were validated by heat shock treatment and subsequent spotting on TSAB agar plates.
Acid and osmotic stress assays.Frozen pneumococcal stocks were thawed at room temperature and serially diluted to 10−5 in THY medium, and 10 μl was spotted onto TSAB plates adjusted to pH 6.0 either by addition of HCl (acid stress) or 250 mM NaCl (osmotic stress). Dilutions were spotted onto TSAB plates with no additions as a control. CFU were calculated after overnight growth at 37°C in 5% CO2.
Whole-genome sequencing of heat shock suppressor mutants.Nine HSR mutants as well as the Δpde1 Δpde2 strain were analyzed by whole-genome sequencing. Bacterial stocks were grown in 10 ml THY medium until the late log phase. Cultures were harvested by centrifugation at 3,500 rpm for 10 min at room temperature. Pellets were washed once in 1 ml TES buffer (100 mM Tris-HCl, 20 mM EDTA, 0.8% SDS) at pH 8.0 before resuspension in 630 μl TES. Next, 70 μl 1% deoxycholic acid sodium salt, 37.5 μl 10 mg·ml−1 RNase A, and 65 μl 20 mg·ml−1 pronase were added, and the mixture was incubated at 37°C until it reached complete cell lysis. Genomic DNA was purified from this mixture by phenol-chloroform extraction and dissolved in a buffer containing 10 mM Tris-HCl (pH 7.5) and 1 mM EDTA (TE) (pH 8.0). Whole-genome sequencing and bioinformatics analysis were performed at the Applied Genomic Technologies Core and the Bioinformatics Core at Wadsworth Center, New York State Department of Health (Albany, NY).
Construction of pneumococcal mutants.S. pneumoniae mutants were generated in an ST581 background by homologous recombination as reported previously (75) using primers listed in Table S2 in the supplemental material. As such, the parent strain ST581 is referred to as the wild type (WT). Generation of Δpde1, Δpde2, Δpde1 Δpde2, ΔtrkH ΔcabP, ΔcabP, and ΔtrkH mutants was reported earlier (17, 36). For deletion mutants, homologous recombination was used to replace the respective ORFs with the Janus cassette, which harbors kanamycin resistance and a dominant rpsL+ allele. Homologous arms were designed to maintain intact genes located upstream and downstream of the targeted ORF. Briefly, homologous upstream and downstream arms were digested with XbaI and XhoI, respectively, and ligated to an XbaI- and XhoI-digested Janus cassette before being transformed. To construct an in-frame mutant, the Janus cassette was removed by transformation with the ligated upstream and downstream arms or with replacement unmarked genes flanked by the upstream and downstream arms. Unmarked mutants were selected as kanamycin sensitive but streptomycin resistant. To delete cabP, trkH, and trkH-cabP in the Δpde1 Δpde2 background, loci, including homologous upstream and downstream arms of the respective gene as well as the Janus cassette, were amplified from previously generated ΔcabP, ΔtrkH, and ΔtrkH ΔcabP mutants, respectively, using the primers listed in Table S2. These fragments were individually transformed into the Δpde1 Δpde2 mutant and selected by kanamycin resistance. To construct the trkH* mutation in the WT (strain ST3326), the ΔtrkH ΔcabP mutant was transformed with the trkH-cabP locus fragment amplified from HSR16 with Pr2976 and Pr2973, which possesses the S69A point mutation. Additionally, this fragment was utilized to transform the Δpde1 Δpde2 ΔtrkH ΔcabP and ΔtrkH mutants to generate trkH* in this background with WT cdaA or cdaA-32. Since cdaA-32 was spontaneously introduced upon transformation of the Δpde1 Δpde2 mutant with ΔtrkH ΔcabP, the Δpde1 Δpde2 cdaA-32 mutant was constructed by replacing ΔtrkH ΔcabP with trkH-cabP WT sequence in the Δpde1 Δpde2 ΔtrkH ΔcabP cdaA-32 mutant. To construct the ΔphoU2 and Δpst2-phoU2 mutant strains, the homologous upstream fragments were amplified with Pr3448 and Pr3449 (phoU2) or Pr3463 and Pr3464 (SPD_1232), respectively, digested with XbaI, and ligated with the XbaI/XhoI-digested Janus cassette and the downstream fragment amplified from Pr3450 and Pr3451 and digested with XhoI. In the Δpst2-phoU2 mutant, the genes SPD_1227 (phoU2) through SPD_1232 were deleted. All mutants were verified by PCR, and point mutations such as trkH* and cdaA-32 were verified by sequencing at GENEWIZ.
Measurement of c-di-AMP concentration by ELISA.Frozen pneumococcal stocks were thawed and grown in 6 ml THY to an OD620 of 0.4. Five milliliters was harvested by centrifugation at 3,500 rpm for 10 min, washed, and resuspended in 500 μl of 50 mM Tris-HCl (pH 8.0). Secreted samples were measured from spent media. E. coli strains were grown overnight in 5 ml of LB broth containing 100 μg·ml−1 ampicillin, as necessary. Five hundred microliters of E. coli cells was harvested by centrifugation and resuspended in 500 μl of 50 mM Tris-HCl (pH 8.0). Cells were lysed by sonication by two 30-s pulses and boiling for 10 min in order to measure intracellular accumulation of c-di-AMP. Soluble lysates or secreted samples were assayed for c-di-AMP through a competitive enzyme-linked immunosorbent assay (ELISA) developed previously (76). The c-di-AMP concentration was normalized to the bacterial density recorded at OD620. Three independent biological replicates were measured for each sample.
Heterologous expression of pneumococcal diadenylate cyclase in E. coli.All plasmids generated in this experiment are listed in Table S3 in the supplemental material. Full-length cdaA was amplified from S. pneumoniae D39 genomic DNA with primers JY348 and JY349. This fragment was digested with XbaI and HindIII and ligated into the vector pGB202, which is a derivative of pBR322 (NEB) that contains the constitutively active promoter for gyrA for expression, to generate pGB217. CdaAV76G was expressed in this plasmid by introducing the T227G mutation with PCR site-directed mutagenesis using primers Pr3330 and Pr3331. CdaA-32 was expressed by amplification with primers JY348 and Pr3514 and cloned between XbaI and BamHI sites in pGB217. E. coli cells harboring individual plasmids were grown overnight for c-di-AMP determinations by ELISA.
Statistical analysis.All data were analyzed by one-way analysis of variance (ANOVA) with Dunnett tests to correct for multiple comparisons to the control or by two-tailed t tests for single comparisons with at least three independent replicates for each sample. Phosphate uptake assays were analyzed by a ratio-paired t test.
ACKNOWLEDGMENTS
We are grateful to Kyle Altman, Jun Yang, and Yuan Liao for expert technical assistance. We thank the Metzger lab members and Joseph Wade (Wadsworth Center, New York State Department of Health) for helpful advice and discussions. We are indebted to the Wadsworth Center Applied Genomic Technologies Core and the Bioinformatics Core Facilities for genomic DNA sequencing and analysis.
This work was partly supported by National Institutes of Health (NIH) grant R01DC006917 to D.W.M. G.B. is a subrecipient of NIH grants R56AI122763 and R35HL135756.
FOOTNOTES
- Received 24 January 2018.
- Accepted 21 February 2018.
- Accepted manuscript posted online 26 February 2018.
- Address correspondence to Guangchun Bai, baig{at}mail.amc.edu.
Citation Zarrella TM, Metzger DW, Bai G. 2018. Stress suppressor screening leads to detection of regulation of cyclic di-AMP homeostasis by a Trk family effector protein in Streptococcus pneumoniae. J Bacteriol 200:e00045-18. https://doi.org/10.1128/JB.00045-18.
For a commentary on this article, see https://doi.org/10.1128/JB.00166-18.
Supplemental material for this article may be found at https://doi.org/10.1128/JB.00045-18.
REFERENCES
- Copyright © 2018 American Society for Microbiology.
