The biofilm inhibitor Carolacton inhibits planktonic growth of virulent pneumococci via a conserved target

New antibacterial compounds, preferentially exploiting novel cellular targets, are urgently needed to fight the increasing resistance of pathogens against conventional antibiotics. Here we demonstrate that Carolacton, a myxobacterial secondary metabolite previously shown to damage Streptococcus mutans biofilms, inhibits planktonic growth of Streptococcus pneumoniae TIGR4 and multidrug-resistant clinical isolates of serotype 19A at nanomolar concentrations. A Carolacton diastereomer is inactive in both streptococci, indicating a highly specific interaction with a conserved cellular target. S. mutans requires the eukaryotic-like serine/threonine protein kinase PknB and the cysteine metabolism regulator CysR for susceptibility to Carolacton, whereas their homologues are not needed in S. pneumoniae, suggesting a specific function for S. mutans biofilms only. A bactericidal effect of Carolacton was observed for S. pneumoniae TIGR4, with a reduction of cell numbers by 3 log units. The clinical pneumonia isolate Sp49 showed immediate growth arrest and cell lysis, suggesting a bacteriolytic effect of Carolacton. Carolacton treatment caused a reduction in membrane potential, but not membrane integrity, and transcriptome analysis revealed compensatory reactions of the cell. Our data show that Carolacton might have potential for treating pneumococcal infections.

which occurs several hours after Carolacton treatment at low pH only 10 , is a secondary effect that might be unique to this species due to the strong acidification of mutans streptococci within mature biofilms 13 .
To investigate these hypotheses and to understand the mode of action of Carolacton in more detail, we studied the response of S. pneumoniae to Carolacton and compared it with that of S. mutans. S. pneumoniae is a ubiquitous colonizer of the nasopharynx 14 whose carriers usually remain asymptomatic. Nevertheless, invasive and non-invasive pneumococcal infections account for almost half a million hospitalizations and approximately 22.000 deaths in the United States annually 15 . A continuous increase in multidrug resistance (MDR, resistant to ≥ 3 classes of antimicrobials) is being reported 16 . Particularly the prevalence of S. pneumoniae strains of serotype 19A has increased drastically from 5% in the mid-1990's to up to 48% in 2008 17 . Serotype 19A isolates have the highest proportion of MDR and extensively drug resistant strains (XDR, resistant to ≥ 5 classes of antimicrobials) 18 . Consequently, the search for novel antimicrobial compounds and novel cellular targets remains of crucial importance, especially with regard to pneumococci.
In this study, we focused on S. pneumoniae TIGR4 as an extensively characterized strain 19 . For the first time, we report complete growth inhibition of planktonically growing cells. We investigated growth inhibition by Carolacton of MDR and XDR clinical isolates of S. pneumoniae classified as serotype 19A. To clarify if S. pneumoniae shares the molecular target of Carolacton with S. mutans, we tested its sensitivity against 9-epi-Carolacton. This Carolacton derivative has the opposite absolute configuration at C-9, the area of macrolactone formation, and is inactive against S. mutans 11 . Cell death of S. mutans biofilm cells has been shown to be dependent on the eukaryotic-like serine/threonine protein kinase (eSTK) PknB 10 and the regulator of cysteine metabolism CysR 20 . Therefore, their homologues in S. pneumoniae TIGR4 were deleted and their response to Carolacton determined. Finally we investigated the transcriptome, including small regulatory RNAs, of Carolacton-treated cultures of S. pneumoniae TIGR4, constructed knock-out mutants of strongly differentially expressed genes and tested their susceptibility. Our data demonstrate that Carolacton acts via the same molecular target in S. mutans and S. pneumoniae. Death of biofilm cells of S. mutans is a late secondary effect. The fast (within 5 min) and distinct transcriptional response to Carolacton treatment represents a compensatory mechanism, since deletion of those genes does not render the mutants insensitive to Carolacton. Finally, the inhibition of growth of S. pneumoniae by Carolacton shows the characteristics of bactericidal and in some cases even bacteriolytic antibiotics.

Results
Carolacton inhibits growth of S. pneumoniae TIGR4 in planktonic culture. Carolacton (0.25 μ g/ml final concentration) was added to cultures of S. pneumoniae TIGR4 prior to the onset of the exponential phase (OD 600 = 0.15, termed "t 0 "). Carolacton-treated cultures grew at the same rate as untreated controls for about 2 h, after which the OD remained constant or decreased (Fig. 1A), suggesting a bactericidal activity 21 . The Carolactontreated cultures typically reached a maximal OD 600 of 0.55-0.60, while the controls grew up to an OD 600 of 1.8. A maximal inhibition of ~67% could be observed after 9 h of growth. Since S. pneumoniae autolyzes after reaching the stationary phase 22 , final cell densities after 24 h were very low and almost the same in treated and untreated cultures. The number of viable cells was determined by counting colony forming units (CFUs). Figure 1B shows that after 3 h of Carolacton treatment, the number of CFUs was reduced by 1 log unit (92.4%) and decreased further by more than 2 log units (99.4%) after 6 h. Since this decrease in CFUs was not accompanied by a reduction in OD 600 , the data indicate reduced cell viability in the presence of Carolacton. For S. pneumoniae TIGR4, the minimal inhibitory concentration (MIC) of Carolacton was 0.06 μ g/ml (Table 1). Cells grown with Carolacton showed a more voluminous phenotype and a fraction of the cells showed a moderate but significant increase in size (Fig. 1C). On average, cell length increased from 1.62 ± 0.4 μ m to 2.08 ± 0.6 μ m, while cell width increased from 0.87 ± 0.12 μ m to 0.94 ± 0.11 μ m. Even a 100-fold increase (0.025-2.5 μ g/ml) of the Carolacton concentration did not increase the inhibition of growth considerably ( Supplementary Fig. S1). This is in full accordance with Kunze et al. who reported a concentration-independent effect of Carolacton on death of S. mutans biofilms over the same concentration range 8 .
Carolacton inhibits planktonic growth of S. pneumoniae serotype 19A clinical isolates. To investigate the effectiveness of Carolacton treatment on S. pneumoniae strains found in pneumococcal disease, three different recent clinical isolates of S. pneumoniae, representing drug resistant serotype 19A strains ( Table 2,  Supplementary Table S1), were subjected to Carolacton susceptibility testing. Isolate Sp49 had been obtained from the blood of a septic pneumonia patient and was resistant against two classes of antibiotics. Isolate Sp61 and Sp64 were XDR (resistant to ≥ 5 classes of antibiotics); they were isolated from the thorax drainage of a pleuritis patient and by swab from an ear canal of a non-symptomatic patient, respectively. The growth of all three clinical isolates was inhibited by Carolacton, but to a widely varying extent (Fig. 2). The clinical pneumonia isolate (Sp49) was most sensitive and showed a different behaviour from TIGR4: it started to die immediately and after 2 h its OD was lower than at t 0 , suggesting a bacteriolytic effect of Carolacton ( Fig. 2A). A maximum inhibition of 94.9% could be observed after approximately 6-7 h of growth, this was the same time point at which the control culture of Sp49 typically reached its peak (OD 600 ~1.2). For the clinical pleuritis isolate (Sp61) and the non-invasive mid-ear serotype 19A isolate (Sp64), the period of uninhibited growth was longer than for the TIGR4 strain, and they reached higher final culture densities than TIGR4. The pleuritis isolate Sp61 showed a maximal inhibition of growth of about 45% after 6 h (Fig. 2B). For the non-invasive mid-ear isolate Sp64, the maximal inhibition was less than that detected for the pleuritic isolate, reaching a maximum inhibition of 23%. Moreover, the inhibition peak was retarded and observed after 8 h (Fig. 2C). Carolacton had a strong lytic effect on clinical isolate Sp49; the MIC value determined for this isolate was ≤ 0.03 μ g/ml and thus lower than for TIGR4. Drug-resistant isolates Sp61 and Sp64, in contrast, were highly tolerant of Carolacton and were only inhibited at 64 μ g/ml Carolacton (Table 1). These results indicate that Carolacton inhibits growth of clinical S. pneumoniae isolates with varying efficiency. The degree of inhibition was not affected by the change of serotype from TIGR4 (serotype 4) to the Scientific RepoRts | 6:29677 | DOI: 10.1038/srep29677 clinical isolates (serotype 19A). Interestingly, growth inhibition was weaker in the pleuritis isolate Sp61 and in the non-invasive isolate Sp64, which elicit multiple resistances to antibiotics compared to the pneumonia isolate.
The interaction of Carolacton with its cellular target is highly stereo-specific. The Carolacton derivative 9(R)-Carolacton (epi-Carolacton) is a diastereomer of Carolacton exhibiting an inversion at C-9 [(S)→ (R)] (Fig. 3A). Although the structural changes are minute, epi-Carolacton is biologically inactive against S. mutans in biofilms 11 . Likewise, 0.25 μ g/ml epi-Carolacton, a concentration which is generally applied in our growth experiments with streptococci, did not inhibit growth of S. pneumoniae TIGR4 (Fig. 3B). Thus the interaction between Carolacton and its molecular target must be highly specific with respect to the stereochemistry  at position 9 of the molecule, heavily influencing the topology of the macrocycle. Due to the stereospecificity of this interaction we suggest that the cellular target of Carolacton is likely to be identical between S. mutans and S. pneumoniae, in spite of the differences of their physiological responses.
The serine/threonine protein kinase StkP and the regulator CysR are not required for growth inhibition by Carolacton. Gene deletion of the signalling kinase PknB induces Carolacton resistance of S. mutans biofilms 10 . The S. pneumoniae genome encodes stkP (sp_1732) as corresponding homologue to S. mutans pknB. Therefore, the role of the homologue of PknB in S. pneumoniae, StkP, in Carolacton susceptibility of S. pneumoniae TIGR4 was tested. Since the molecular target of Carolacton appears to be highly conserved in those two organisms, this analysis will clarify whether a bacterial eukaryotic-like serine/threonine protein kinase (eSTK) is the direct molecular target of Carolacton. Due to the strong indications of eSTKs as the primary target of Carolacton 10 , in addition to PCR amplification and DNA-sequencing, the absence of the StkP protein in the Δ stkP isogenic mutant was confirmed by immunoblotting using polyclonal anti-StkP antibodies 23 (Fig. 4A). The StkP deletion strain showed formation of elongated cells and multiple unconstricted septa, as reported previously 23 . Its growth was highly impaired; nevertheless, Carolacton treatment led to growth inhibition (Fig. 4B). The data show that the Δ stkP TIGR4 strain is susceptible to Carolacton, which is in sharp contrast to the findings described for S. mutans 10 . Deletion of the cysteine metabolism regulator CysR also resulted in a Carolacton-insensitive mutant in S. mutans 20 . The CysR homologue (SP_0927) was deleted in S. pneumoniae and tested for susceptibility to Carolacton. Similar to the Δ stkP strain, growth of the Δ cysR mutant was strongly inhibited by Carolacton (Fig. 4C). Thus, the pneumococcal homologues to both proteins mediating the susceptibility to Carolacton in S. mutans, StkP and CysR, are most likely not directly involved in the Carolacton response in S. pneumoniae, in spite of the presumably conserved molecular target of Carolacton in those two species. We conclude that PknB and CysR are involved in the mechanisms of membrane damage at low pH in S. mutans, which is a late secondary response to Carolacton treatment. In order to confirm this hypothesis, the effect of Carolacton on membrane integrity and membrane potential of S. pneumoniae TIGR4 was analysed.

Effect of Carolacton on cytoplasmic membrane integrity and membrane potential in S. pneumoniae.
For S. mutans it has been reported that Carolacton leads to death of biofilm cells at low pH by disturbing the integrity of the plasma membrane and thereby causing leakage of cytoplasmic proteins and DNA 10 . Membrane damage of S. pneumoniae cells was similarly determined by LIVE/DEAD staining during exponential and stationary phase, 3 h and 6 h after addition of Carolacton, respectively. Heat-killed S. pneumoniae cells were included as a control and a calibration curve was recorded using flow cytometry ( Supplementary Fig. S2). No indication of S. pneumoniae cell death by Carolacton treatment was detected ( Supplementary Fig. S3). Controls and Carolacton-treated cultures both contained a maximum of 14% cells with defective plasma membranes (Fig. 5A).
We then used staining with the fluorescent dye 3,3-Diethyloxacarbocyanine iodide (DiOC 2 (3)) to determine changes in the MP upon Carolacton treatment. Since the measured fluorescence intensities are directly influenced by the size of the stained particles, and Carolacton treatment affects the cell size, a size-independent ratiometric technique was applied 24 . Carolacton treatment caused a time-dependent depolarization of the membrane potential (Fig. 5B, Supplementary Fig. S4). Approximately 48% of the cells showed a membrane depolarization, indicated by a reduction of the red/green fluorescent ratio, after 3 h. Membrane depolarization continued during the next 3 h, though not as fast as during the first time interval. After 6 h, the red/green fluorescent ratio dropped to ~34% of that measured for an exponentially growing control after 3 h, and was reduced to ~56% when compared to the control at the same time point. These results indicate a Carolacton-dependent disturbance of the membrane potential during growth that is not the result of a porous plasma membrane.  Transcriptional response of S. pneumoniae TIGR4 to Carolacton. We analysed the transcriptome of S. pneumoniae TIGR4 cells growing with Carolacton starting 5 min after addition of the drug, until 3 h after Carolacton addition. Total RNA was fractionated into two size classes: > 200 nt and ≤ 200 nt, and differential regulation of mRNAs (30-4776 nt), small RNAs (37-486 nt) and tRNAs (71-90 nt) was assessed individually. We noted a high number of reads for the previously identified housekeeping sRNAs 6S RNA (194 nt) 25 and RNase P (391 nt)/tmRNA (348 nt) 26,27 in the samples containing small transcripts (≤ 200 nt) or in the samples containing large transcripts (> 200 nt), respectively, confirming the applicability of our sequencing approach for reliable separation of transcripts according to length (Supplementary Dataset 1). Since differential transcription after Carolacton treatment was exclusively limited to the later time points (t 120 /t 180 ), we chose to only sequence three time points (t 0 , t 5 and t 180 ) for analysis of short small RNAs/tRNAs. The RNA-seq data of biological replicates showed a very high Pearson correlation between 0.93 for small regulatory RNAs/tRNAs and 0.99 for large regulatory RNAs/mRNAs. None of the 58 tRNAs of S. pneumoniae TIGR4 were significantly differentially expressed after Carolacton treatment ( Supplementary Fig. S5, Supplementary Dataset 1).  Fig. S5). The highest upregulated small sRNA trn0057 (log 2 FC~3.9) is located downstream of the sp_0119-sp_0120 operon and its function is unknown. Ten long sRNAs were differentially regulated at t 120 and t 180 ( Supplementary Fig. S5) of which only two (R12 and F20) have been investigated and were shown to be important for virulence in mice 28 .
Two transcripts smaller than 200 nt were differentially regulated, a type II DNA modification methyltransferase (sp_0569) and the 50S ribosomal protein L28 (sp_0441) (Supplementary Fig. S5). 39 transcripts larger than 200 nt were strongly differentially expressed (log 2 FC ≥ ± 2 and FDR < 0.01 for at least one time point) (Fig. 6), 26 of which are encoded within 7 operons on the TIGR4 chromosome (Table 3). Differential expression of sp_0119 and sp_0120 was detected already 5 min after addition of Carolacton and continued to increase during the experiment to a final log 2 FC of 4.4 and 4.3 after 180 min, respectively. Sp_0119 codes for a MutT/Nudix family protein and sp_0120 encodes the glucose-inhibited division protein A (GidA).
Influence of sp_0119 and sp_0120 on susceptibility of S. pneumoniae TIGR4 to Carolacton.
Since the gene products of both sp_0119 and sp_0120 are involved in central cellular processes we wondered whether an uncontrolled transcriptional upregulation of this operon could cause the growth inhibitory effect of Carolacton. Therefore, single deletion mutants of sp_0119 and sp_0120 and a Δ sp_0119-sp_0120 double mutant were constructed in S. pneumoniae TIGR4. Growth of all three mutant strains was still significantly inhibited by Carolacton treatment (Supplementary Fig. S7). Comparison of the transcriptome of S. mutans UA159 and S. pneumoniae TIGR4 in response to Carolacton. Until now, the only other organism for which the influence of Carolacton has been studied in more detail is the human oral pathogen S. mutans. Therefore, the RNA-seq data obtained in this study were compared to microarray data previously obtained for Carolacton-treated S. mutans biofilm cells 10 . 22.8% of all genes were differentially transcribed at 180 min in S. pneumoniae at a log 2 FC of ≥ ± 0.8 ( Supplementary Fig. S6), which had previously been used for S. mutans 10 . At this cut-off, the effect of Carolacton was highly pleiotropic. Supplementary Fig. S8 shows a direct comparison between log 2 FC of the most strongly differentially expressed genes in S. mutans compared to their homologues in S. pneumoniae. Overall, the patterns of transcriptional changes were similar for some, but not all genes, and the timing and extent of the responses shows differences (Supplementary Fig. S8). For example, upregulation of the ComDE two-component system (TCS) for genetic competence is observed in both species, but stronger and earlier in S. pneumoniae, and might represent a general stress response 29 . The most conspicuous transcriptional response in S. mutans was the instantaneous strong upregulation of genes for pyrimidine biosynthesis. In S. pneumoniae, transcription of these genes showed a different pattern, indicating that their regulation in S. mutans might be related to the membrane damage occurring at later stages of biofilm growth. In S. mutans, downregulation of sortase A was observed. In S. pneumoniae TIGR4, neither sortase A (srtA, sp_1218) nor its three sortase paralogs (srtBCD, sp_0466-sp_0468) 30 were differentially transcribed. Also in contrast to S. mutans, the VicKRX TCS was not differentially expressed upon Carolacton Inhibition of the TIGR4 wild type strain (dotted lines), used as controls during both assays (Carolactontreated = black triangles, control = white triangles), is displayed as a semi-transparent plot in (A) and (B). All growth curves present the average of three independent biological replicates. treatment in S. pneumoniae. Consequently, the VicKRX regulon also showed a different pattern: only transcripts of the cell shape proteins MreC and MreD were upregulated, although earlier and stronger than in S. mutans. In S. mutans, cell wall metabolism-related genes were upregulated, while the respective homologues in S. pneumoniae TIGR4 were not significantly differentially expressed.
Interestingly, the gidA homologue of S. mutans (smu_2141) was among the most strongly upregulated transcripts like in S. pneumoniae, although its upregulation was weaker and occurred later than in the pneumococcus ( Supplementary Fig. S8).
Taken together, the transcriptional response of S. mutans and S. pneumoniae to Carolacton appears to be a pleiotropic stress response, different between species, and not a specific and focused transcriptional response as assumed after studies with S. mutans before 10 .

Discussion
For the first time we report on the physiological effect of Carolacton on the clinically relevant human pathogen S. pneumoniae. Growth of planktonic cultures of S. pneumoniae is already inhibited by Carolacton at a concentration as low as 0.025 μ g/ml ( Supplementary Fig. S1). Growth inhibition starts about 2 h after application of the drug; at later time points, the optical density remains constant, but the number of viable cells continues to decrease. Such behaviour is typically observed when treating pneumococci with bactericidal antibiotics 21 . Growth inhibition by Carolacton is concentration-independent, which could indicate that the primary cellular target of Carolacton is easily saturated and thus of low abundance in cells.
Here we show that growth inhibition by Carolacton occurs for clinical isolates as well as for the virulent laboratory strain TIGR4 and thus appears to be independent of the serotype. Interestingly, the susceptibility to Carolacton was related to the drug resistance of the pathogen. The sepsis isolate Sp49 (which is resistant to two classes of antibiotics only) was instantly killed by Carolacton, while the multidrug-resistant isolates Sp61 and Sp64 were inhibited to a lower degree than the TIGR4 strain. Accordingly, also the MIC values of Carolacton differed notably between S. pneumoniae TIGR4 and among the clinical isolates. Carolacton has a potent inhibitory effect on S. pneumoniae TIGR4 and Sp49 with MIC values in the range of MICs of classical antibiotics used for the therapy of pneumococcal disease, such as e.g. ampicillin, clarithromycin or rifampicin 31 . The exact MIC of Sp49 remains to be determined and could be even lower than 0.03 μ g/ml. Multidrug-resistant isolates Sp61 and Sp64 can tolerate and grow in the presence of high concentrations of Carolacton. Multidrug resistance is often mediated by increased expression or activity of multidrug efflux pumps. Sp61 and Sp64 are resistant to macrolides (clarithromycin, telithromycin) 32 and fluoroquinolones (levofloxacin) 33 , which is mediated by efflux in the pneumococcus. Carolacton is structurally similar to the macrolide class of natural compounds and might be exported in a similar way. Unspecific efflux of Carolacton over time, as a by-product of resistance to macrolides and fluoroquinolones, could explain why the MICs of Sp61 and Sp64 are much higher than that of S. pneumoniae TIGR4. This is in accordance with the observation that E. coli is insensitive to Carolacton, but the growth of E. coli TolC, which lacks a multidrug efflux pump, is completely inhibited 5 . These findings are important for future A Carolacton derivative with an inverted stereocentre at C-9 (epi-Carolacton) was completely inactive against both S. mutans and S. pneumoniae, as well as against several oral streptococci 11 . All other chemical modifications of the side-chain or the lactone ring tested so far were also inactive 11 . This indicates a highly stereospecific interaction of Carolacton with its molecular target and led us to hypothesise that the target could be highly conserved and shared between S. pneumoniae and S. mutans. It appears unlikely that a switch of the stereochemistry at C-9 could completely abolish interaction with multiple less conserved targets in both species. If Carolacton would interact with several weakly conserved targets, we would expect a remaining activity of epi-Carolacton in one of the tested species, which is absent as shown here and previously 11 . A pattern similar to the one described for epi-Carolacton here was found for the antibiotic fosfomycin, which targets the conserved bacterial cell wall hydrolase MurA 34 , but its stereoisomer lacks antimicrobial activity 35 . Consequently we investigated the roles of PknB and CysR, which both had been shown to be essential for the damage of biofilm cells of S. mutans at low pH 10,20 , for Carolacton susceptibility in S. pneumoniae. Our data indicate that neither StkP nor CysR are required for susceptibility of S. pneumoniae to Carolacton and they thus cannot be its molecular target if this is conserved in both streptococci. Another indication that Carolacton acts independently of StkP is the potent activity of Carolacton against E. coli TolC, as E. coli does not possess StkP homologues 36 . StkP and CysR are apparently exclusively important for S. mutans at low pH in mature biofilms, which is a unique environment fundamentally different from planktonically growing S. pneumoniae cultures.   damage is usually paralleled by a loss of membrane potential (MP) 37 . Treatment with antimicrobial agents such as streptomycin, reutericyclin 38 or ampicillin 37 can decrease the MP and thereby interfere with the localization of cell division proteins 39 . By fluorescent staining of Carolacton-treated cells with DiOC 2 (3), we detected a continuous decrease in MP over time. When assessing the integrity of the plasma membrane, as a possible cause for MP reduction, no damage could be detected. This is in agreement with Kunze et al. who reported no changes in LIVE/ DEAD-stained S. mutans cells during planktonic growth with Carolacton 8 . Thus, Carolacton treatment leaves the plasma membrane intact in both S. mutans and S. pneumoniae during planktonic growth. To identify components of a possible conserved response in S. mutans and S. pneumoniae, RNA-sequencing was carried out. Among differentially regulated sRNAs, sspn1731.1 and sspn1732.1 are strongly downregulated at late time points. Both contain a T-box motif that is known to stimulate anti-termination of the tyrS gene in Bacillus subtilis by interaction with the acceptor end of tRNA(Tyr) 40 . T-box leader sequences of non-coding RNAs are part of a negative feedback loop controlling expression of aminoacyl-tRNA synthetase genes 41 and components of the amino acid biosynthesis machinery 42 . Downregulation of sspn1731.1 and sspn1732.1 may present another compensatory mechanism to Carolacton by increasing expression, for example of aminoacyl-tRNA-synthetases. Among differentially regulated mRNAs, an immediate strong upregulation of the sp_0119-sp_0120 operon was identified. Both encoded proteins have a global regulatory role in S. pneumoniae. The sp_0119 gene encodes a MutT/Nudix family-like protein. A homologous protein in E. coli (MutT) functions as an anti-mutator protein, by maintaining replication fidelity through hydrolysis of mutagenic 8-oxo-dGTP 43 . Other MutT/Nudix family-like proteins in E. coli similarly protect the cell by degradation of toxic nucleotide derivatives 44 and have been described as "housecleaning" enzymes 45 . Thus, it can be assumed that SP_0119 likewise plays a role in protecting the pneumococcus by detoxification of harmful metabolites. GidA (SP_0120) is a highly conserved protein; homologues are present in Proteobacteria and Firmicutes and influence protein biosynthesis by altering mRNA codon recognition through modification of tRNAs at the wobble position 46 . GidA modulates the virulence of a large number of pathogens 47 . Consequently, a Salmonella gidA mutant strain is currently being investigated for use in a live-attenuated vaccine 48 . As an important virulence factor, also in streptococci 49 , it is conceivable that by altering the transcription of gidA, Carolacton may also interfere with virulence of S. pneumoniae. Expression of virulence factors is tightly controlled in group A streptococci 50 and upregulated transcription of virulence factors can have a negative effect on pathogenicity, as shown for Listeria monocytogenes 51 . The investigation of compounds that are aiming at virulence factors in order to render pathogens harmless, instead of targeting essential cellular components, hold enormous possibilities 52 . The use of anti-virulence drugs could even counteract resistance development in bacteria altogether 53 . If S. pneumoniae cells show an altered virulence after treatment with Carolacton needs to be explored further.
Growth of single and double deletion mutants of sp_0119 and sp_0120 was still impaired by Carolacton, thus neither SP_0119 nor SP_0120 qualify as its molecular target. We hypothesise that these proteins mediate a compensatory response to Carolacton treatment. This is further confirmed by comparison of the transcriptional response of S. mutans and S. pneumoniae, which suggests that these responses likely show species-specific stress reactions or compensatory mechanisms and indicate that the effect of Carolacton cannot be compensated by differential regulation of a single specific gene set. For example, competence development, e.g. by upregulating transcription of comDE, is a direct reaction to stress in bacteria, such as streptococci, that lack a classical SOS response 54 . An upregulation of the VicKRX operon and VicKRX-regulated genes, as seen in S. mutans, possibly presents a direct response to disturbance of cell wall homeostasis 55 . The proteins MreC and MreD are located at the equators and septa of dividing S. pneumoniae cells and direct peripheral peptidoglycan synthesis 56 . Their increased transcription after Carolacton treatment could be interpreted as an attempt to compensate defects in peptidoglycan synthesis and cell division.
The data show that Carolacton holds promise as a bactericidal antibiotic. The extent of growth inhibition observed after Carolacton treatment in S. pneumoniae is variable, does not depend on the serotype and can even be bacteriolytic. This calls for further investigation of the primary mode of action of Carolacton and demonstrates the importance to discover the direct molecular target as a pre-requisite for rational drug design.

Methods
Additional methods can be found in the Supplementary Information. Bacterial strains and growth conditions. Bacterial strains used in this study are listed in Table 2.
S. pneumoniae TIGR4 serotype 4 (ATCC BAA-334), used for growth experiments in liquid medium, was routinely grown on Columbia blood agar plates containing 5% sheep blood (BD Biosciences) overnight (o/n) at 37 °C and 5% CO 2. Cells were then scraped from plate and used to inoculate Todd-Hewitt-Broth + 1% (w/v) yeast extract (THBY, BD Biosciences) liquid medium, starting from an optical density at 600 nm (OD 600 ) of 0.1. After the initial culture had reached an OD 600 of 0.15, it was split into equal volumes and Carolacton or 9(R) epi-Carolacton 11 were added to a final concentration of 0.25 μ g/ml, if not stated otherwise. Carolacton was solved in methanol and stored 1000-fold concentrated (5.3 mM) in small aliquots in glass vials at − 20 °C in the dark. Epi-Carolacton was solved in dimethyl sulfoxide (DMSO) and also stored in concentrated small aliquots (2 mM) at − 20 °C in the dark.
Clinical S. pneumoniae isolates and information about clinical background data, serotype and antimicrobial susceptibility were obtained from Dr. Mark van der Linden from the National Reference Laboratory (NRZ) on Streptococcal Disease in Aachen (Germany). All clinical strains were isolated between 5/2008 and 4/2009. For cryo-conservation, S. pneumoniae strains were grown in THBY to early exponential phase (OD 600~0 .25) and immediately frozen at − 80 °C in 10% (v/v) glycerol.
Growth inhibition was calculated as follows: Scientific RepoRts | 6:29677 | DOI: 10.1038/srep29677 Determination of MIC values. S. pneumoniae was grown on Columbia blood agar plates o/n as described before. Minimum inhibitory concentration (MIC) values of Carolacton were determined by two-fold serial microdilution and incubation at 37 °C and 5% CO 2 for 18-20 h, as described for the pneumococcus previously 31 .
The dilution range of Carolacton tested was 0.03-128 μ g/ml. The MIC was determined as the lowest concentration of Carolacton that did not yield visible bacterial growth. The cell number of the initial inoculum was 5 × 10 5 CFU/ml as determined by serial dilution and plating. Ampicillin, erythromycin, tetracycline and rifampicin were included as controls during MIC determination for S. pneumoniae TIGR4, confirming target MIC values commonly reported for susceptible pneumococci 31 . MICs were confirmed by at least two independent measurements (with two replicates). of competence by addition of competence stimulating peptide 2 (CSP-2) as described previously 57 . Briefly, THBY (pH 6.8) supplemented with 1 mM CaCl 2 and 0.2% (w/v) bovine serum albumin (BSA) was inoculated with exponentially growing cells from liquid culture in THBY (OD 600~0 .5) to an OD 600 of 0.005. After reaching the early log phase (OD 600~0 .15), 1 ml culture was harvested by centrifugation and the cells resuspended in 1 ml THBY (pH 8.0) supplemented with 1 mM CaCl 2 , 0.2% (w/v) BSA and 400 ng/ml CSP-2. Transformation was achieved by addition of 10 μ l of a ligation mix and static incubation at 37 °C for 2 h. Transformants were selected on THBY agar plates containing 0.5 μ g/ml erythromycin. Correct integration was checked by PCR and Sanger DNA sequencing.

Verification of the absence of StkP in the TIGR4ΔstkP deletion mutant by immunoblotting.
In addition to DNA sequencing, the absence of StkP of JD01 was confirmed by immunoblotting. First, a defined volume (V[μ l] = 300/OD 600 ) of S. pneumoniae TIGR4 wild type and TIGR4Δ stkP cells in THBY was boiled in SDS loading dye (95 °C, 15 min) and separated via 12% (v/v) SDS PAGE according to 59 . Proteins were blotted to a PVDF membrane (Immobilion-P, Milipore) using a semi-dry blotting system (Sigma-Aldrich) and blocked with 5% (w/v) BSA in TBS-T at room temperature (RT) for 1 h. Incubation with the primary antibody, rabbit anti-StkP, was carried out by using a 1:50.000 dilution in 1% (w/v) BSA/TBS-T o/n. Pneumococcal α -enolase (α -eno) served as a loading control 60 . The α -eno antibody was used in a 1:500 dilution in 1% (w/v) BSA/TBS-T at RT o/n. Alkaline phosphatase-coupled goat anti-rabbit IgG (Sigma-Aldrich) diluted 1:10 000 in TBS-T was applied for 1 h at RT for detection of the primary antibody. Alkaline phosphatase activity was visualized colorimetrically using SIGMA FAST BCIP/NBT tablets according to the manufacturer's instructions (Sigma-Aldrich).

RNA-seq data analysis.
Processing of RNA-seq data is described in Supplementary Information. All analysed small RNAs are documented in Supplementary Dataset S3. False discovery rate (FDR)-adjusted P values were calculated according to 61 . FDR values of < 0.01 were considered significant. For visualization of these results heat maps were generated, taking into account only genes that, in addition to a significant FDR, showed a log 2 -fold change (log 2 FC) of transcription of ≥ ± 2 (or ≥ ± 0.8 for small RNAs) at least at one point during a time course. For this, the log 2 -fold values of transcript abundance obtained in edgeR were used in combination with the heatmap. 2 function of the R package gplots (v2.15.0) 62 . Raw and processed RNA-seq data have been deposited in NCBI's Gene Expression Omnibus (GEO) database 63  For staining with DiOC 2 (3), cells were diluted from culture to approx. 10 7 cells/ml and stained with 30 μ M DiOC 2 (3) in the dark for 30 min. In order to assess bacterial membrane integrity, cells were separately co-stained with the fluorescent dyes Syto 9 and propidium iodide (PI) using the LIVE/DEAD BacLight Bacterial Viability Kit (Life Technologies). For Syto 9/PI staining, cells from a culture grown with 0.25 μ g/ml Carolacton and from a control were harvested by centrifugation and washed in 0.85% (w/v) NaCl. Approx. 5 × 10 6 cells/ml were stained with an equal volume 2× staining solution (10 μ M Syto 9, 60 μ M PI) at RT in the dark for 30 min. Heat-inactivated cells were used as a control in both staining experiments. Heat-killing of cells was achieved by incubation at 80 °C and shaking (1,400 rpm) for 30 min. Microscopy and quantification of cell size. S. pneumoniae TIGR4 cells growing in THBY were washed with 0.85% (w/v) NaCl solution and mounted to a microscope slide coated with a thin layer of 1% (w/v) agarose. Phase-contrast microscopic analysis of S. pneumoniae was performed using an Olympus BX60 microscope equipped with an Olympus DP50 color camera (5 Mpixel) color view II camera and a 100x magnification oil immersion objective. Pictures and measurements were taken with the Olympus cellSens (v1.6) microscopy software. The length of pneumococcal cells was measured along the length axis. Cell width was determined as the diameter along the equatorial plane, perpendicular to the length axis of the cell.
Quantification of colony forming units. For determination of colony forming units (CFUs), cells were treated with Carolacton (0.25 μ g/ml) prior to exponential growth. Ten-fold serial dilutions of cultures were prepared in 0.85% (w/v) NaCl solution (pre-warmed to 37 °C) after 3 h and 6 h of growth. A total volume of 100 μ l from each dilution was then plated onto pre-warmed (37 °C) THBY agar plates, which were incubated at 37 °C and 5% CO 2 o/n.

Statistical analysis.
Comparisons between two groups at a single time were performed with unpaired two-tailed (non-parametric) Mann-Whitney U test 64 or with Student's t-tests for two-tailed distribution (with a critical P value of 0.05). The data were analysed for normality by Shapiro-Wilk test 65 . Statistical analyses were performed with the statistics options of SigmaPlot (v13.0) software (Systat Software Inc.).