Antibiotic Stress Induces Genetic Transformability in the Human Pathogen Streptococcus pneumoniae

See allHide authors and affiliations

Science  07 Jul 2006:
Vol. 313, Issue 5783, pp. 89-92
DOI: 10.1126/science.1127912


Natural transformation is a widespread mechanism for genetic exchange in bacteria. Aminoglycoside and fluoroquinolone antibiotics, as well as mitomycin C, a DNA-damaging agent, induced transformation in Streptococcus pneumoniae. This induction required an intact competence regulatory cascade. Furthermore, mitomycin C induction of recA was strictly dependent on the development of competence. In response to antibiotic stress, S. pneumoniae, which lacks an SOS-like system, exhibited genetic transformation. The design of antibiotherapy should take into consideration this potential of a major human pathogen to increase its rate of genetic exchange in response to antibiotics.

Bacterial transformation, originally discovered in the human pathogen S. pneumoniae (1), relies on a process that is inherent to the species, is independent of extrachromosomal elements, and can be considered the only programmed mechanism for generalized genetic exchange in bacteria. It allows the uptake and integration of exogenous DNA in the recipient genome and is considered to be a form of parasexuality (2). Transformation is believed to contribute to the genetic plasticity of S. pneumoniae and to play a central role in the adaptation of this pathogen to host defenses (3). We sought to establish whether transformation is induced in response to antibiotic stress.

In S. pneumoniae, competence for genetic transformation is a transient physiological state allowing efficient DNA uptake (4) and a previously unrecognized capacity to kill noncompetent cells (5, 6), a phenomenon referred to as pneumococcal fratricide (6) or sobrinicide (7). The development of competence requires transcriptional activation of the com regulon, which comprises 105 to 124 genes (8, 9), including recA (10). The RecA protein plays a key role in transformation by catalyzing homologous recombination between the internalized DNA and the recipient genome. The com regulon is induced when a competence-stimulating peptide (CSP), encoded by comC and exported through a dedicated secretion apparatus ComAB, accumulates in the medium and stimulates its receptor, the membrane-bound histidine kinase ComD (4). It is assumed that ComD then autophosphorylates and transphosphorylates its cognate response regulator ComE (11), which in turn activates the expression of the so-called early com genes (4), including comAB, comCDE, and comX. The latter encodes an alternative sigma factor σX (12), which most probably recognizes a sequence (TACGAATA, hereafter called Pcin) conserved in the putative promoter regions of the late com genes (4, 9). The early control of competence induction is not yet fully understood. It was first suggested that competence induction relies simply on passive CSP accumulation, but we favor an alternative model in which CSP production could be temporarily increased in response to changes in environmental conditions (4, 13). We further propose that competence in S. pneumoniae is a general stress response, playing a role similar to that of the SOS response in Escherichia coli (4).

We tested this hypothesis by investigating the effect of mitomycin C, a DNA-damaging agent known to induce the SOS response, on the com regulon. To monitor competence induction (14), we used transcriptional fusions with the luc gene from the firefly (Photinus pyralis) encoding luciferase, the activity of which can be monitored directly in living S. pneumoniae cells (14). We first used a fusion of luc with ssbB, a representative of the late com genes (8, 9), which encodes a single-stranded DNA–binding protein (15). Luciferase activity was monitored during growth with various concentrations of mitomycin C. Expression of the reporter was stimulated by exposure to 25 to 60 ng ml–1 of mitomycin C (Fig. 1A and fig. S1A), indicating that ssbB was induced when cells were grown in the presence of the DNA-damaging agent. No induction of the same fusion was detected in a comA mutant background [that is, in a strain unable to export CSP and therefore to develop spontaneous competence (Fig. 1B)], demonstrating that an intact competence regulatory cascade was required for induction by mitomycin C. These data established that a DNA-damaging agent induced the com regulon of S. pneumoniae.

Fig. 1.

Mitomycin C (MC) induction of ssbB and recA. (A and B) Luciferase activity expressed in relative luminescence units (RLU)/optical density (OD) (triangles) and OD492 (squares) of cultures in C+Y medium of (A) strain R895 (ssbB::luc) and (B) strain R1313 (ssbB::luc, comA) without and with mitomycin C (40 ng ml–1). Curves of luciferase activity [with (red triangles) and without (gray triangles) mitomycin C] and OD [with (black squares) and without (gray squares) mitomycin C] represent compilations of data from 19 and 32 replicate cultures, respectively in (A) and (B). Standard deviations are indicated for luciferase activities only. (C) Structure of the recA operon with the recA::luc fusion generated by the integration of plasmid pR432. Plasmid sequences are not drawn to scale. Pcin and Pa are indicated by the small branched arrows. (D and E) Luciferase activity (triangles) and OD492 (squares) of cultures in C+Y medium of strain R1624 (recA::luc, ΔcomC). In (D), solid blue triangles and solid black squares indicate the presence of CSP (100 ng ml–1), whereas open gray symbols indicate the absence of CSP. In (E), symbols indicate the following concentrations of mitomycin C: solid gray symbols (0 ng ml–1), red triangles and open gray squares (25 ng ml–1), open black symbols (100 ng ml–1), and solid black symbols (400 ng ml–1). Arrows indicate the addition of CSP or mitomycin C after 70 min of incubation.

Mitomycin C is known to trigger the SOS response in E. coli. DNA damage caused by mitomycin C blocks the replication fork, generating a single-stranded DNA region to which RecA binds to form a nucleoprotein filament (RecA*) (16). The coprotease activity of RecA* then catalyzes the self-cleavage of the SOS repressor LexA (17), which leads to induction of the SOS genes, including recA. In S. pneumoniae, the recA gene has been shown to be expressed from two promoters: Pcin, which generates a 5.7-kb-long transcript in competent cells, and Pa, a σA promoter that directs the synthesis of a 4.3-kb-long transcript (18) (Fig. 1C). The competence-specific induction of recA (that is, expression from Pcin) accounts for 95% of the transformation (10). To establish whether mitomycin C could induce recA expression, a recA::luc fusion was constructed (Fig. 1C) and validated by measuring its induction with CSP (Fig. 1D). Mitomycin C was found to activate this fusion with kinetics similar to that of ssbB (fig. S2). The recA::luc construct was introduced in a strain unable to develop spontaneous competence because of the deletion of the entire comC coding region. The DNA-damaging agent did not induce luc expression in this genetic background (Fig. 1E), demonstrating that the induction of recA by mitomycin C occurs only from Pcin and is therefore strictly dependent on the ability of this compound to induce competence. This observation supports the hypothesis that S. pneumoniae, which lacks an SOS-like induction mechanism (7), instead uses the competence regulatory cascade to coordinate a response to mitomycin C. Induction of the com regulon in S. pneumoniae and of the SOS response in E. coli occurs only after prolonged incubation, at ∼2.5 hours (Fig. 1A) and 2 hours, respectively (19). The similar delay of unrelated regulatory mechanisms probably reflects the need for a slow accumulation of inducing lesions in both species.

Several antibiotics are known to induce the SOS response in SOS-proficient bacteria (20, 21). To investigate the parallels between competence induction and the SOS response, we measured luciferase synthesis from the ssbB::luc fusion during growth using a wide range of concentrations of various antibiotics (table S2). Among the protein synthesis inhibitors that were tested, kanamycin and streptomycin triggered competence (Fig. 2, B and C, and fig. S1A), but erythromycin and tetracycline did not (fig. S3, A and B). The fluoroquinolones (norfloxacin, levofloxacin, and moxifloxacin, the latter two of which are used for the treatment of respiratory tract infections), which target type II topoisomerases, DNA gyrase, and topoisomerase IV (22), were found to induce the ssbB::luc fusion (Fig. 2A and fig. S1, B and C). No induction was detected with the DNA gyrase inhibitor novobiocin (fig. S3B), the RNA polymerase inhibitor rifampicin (fig. S3B), the glycopeptide antibiotic vancomycin (fig. S3B), or with the β-lactams ampicillin (fig. S3A) and the third-generation cephalosporin, cefotaxime (fig. S3A). No correlation could be made between the intensity of growth inhibition and the induction of competence (table S2). Similarly to mitomycin C, the induction of ssbB::luc by aminoglycosides and norfloxacin required an intact competence regulatory cascade, because no induction could be detected in a comA mutant (Fig. 2, D to F).

Fig. 2.

Antibiotics induce the competence regulon. (A to F) Luciferase activity (triangles) and OD492 (squares) of cultures of strain R895 (ssbB::luc, wild type) [(A) to (C)], strain R1313 (ssbB::luc, comA) [(D) and (F)], and strain R1047 (E) (ssbB::luc, comA) [(D) to (F)] in C+Y medium with (red and black symbols) and without (gray symbols) antibiotics. Antibiotics were added after 70 minofincubation(arrows) in the following concentrations: 11 μg ml–1 of norfloxacin (Nf) [(A) and (D)], 31.25 μg ml–1 of kanamycin (Kn) [(B) and (E)], and 12.5 μg ml–1 of streptomycin (Sm) [(C) and (F)]. Curves of luciferase activity (with standard deviations) and OD represent compilations of data from 8 cultures [(A) and (D)], 15 cultures [(C) and (F)], or 16 cultures [(B) and (E)] of strains with the respective antibiotics. See Fig. 1 legend for details.

Induction of the com regulon normally allows competent cells to take up and integrate exogenous DNA. To test whether antibiotic-induced competence resulted in a bona fide transformation, transforming DNA was added to a culture treated with streptomycin. For the transformation assay, we selected an intermediate concentration of streptomycin (625 ng ml–1), which did not cause severe killing (Fig. 3B). Chromosomal transformants were readily obtained in the streptomycin-treated culture, whereas no transformants were present in the control culture (Fig. 3, C and D). The addition of CSP (100 ng ml–1) to the streptomycin-treated culture did not further increase ssbB::luc expression (Fig. 3C) or the yield of transformants (Fig. 3D), demonstrating full competence induction by the antibiotic. Chromosomal transformants were also obtained in parallel cultures treated with mitomycin C (60 ng ml–1) and norfloxacin (10 μg ml–1) (fig. S4).

Fig. 3.

Induction of genetic transformation by streptomycin. (A and B) Luciferase activity (A) and growth [(B), OD492 and colony-forming units (CFUs) per milliliter] of cultures of strain R895 with 0 (open squares), 6.25 (open triangles), 12.5 (solid gray triangles), or 25 (solid black squares) μg ml–1 of streptomycin (Sm). Streptomycin was added after 70 min of incubation [arrow in (A)]. See Fig. 1 legend for details. Cell survival was monitored by plating aliquots [dotted lines in (B)]. (C) Luciferase activity of strain R895 grown without (squares) and with (triangles) 6.25 μg ml–1 of streptomycin. CSP (100 ng ml–1) was added (right arrow) after 182 min (solid black triangles). (D) Genetic transformation in aliquots from cultures in (C) taken after 195 min of incubation and mixed with chromosomal DNA carrying a marker conferring resistance to streptomycin (14). The monitoring of transformation with a marker conferring resistance to streptomycin is unrelated to the use of streptomycin to induce competence. The number of streptomycin-resistant chromosomal transformants obtained corresponds to transformation frequencies of 0.65% (Sm) and 0.57% (Sm+CSP).

The induction by mitomycin C and fluoroquinolones of the SOS response in E. coli and of competence in S. pneumoniae suggests that SOS and competence play similar roles in both species. However, the parallel is only partial, because competence was not induced by antibiotics that disrupt cell wall integrity (such as β-lactams), which is contrary to the SOS system in E. coli (21). Aminoglycosides that induce the com regulon (such as kanamycin and streptomycin) did not trigger an SOS response in E. coli but instead induced heat-shock protein expression (23), suggesting that similar stress signals are processed differently in the two species. Mitomycin C and fluoroquinolones may generate a common signal—chromosome replication arrest—owing to the formation of interstrand cross-links (mitomycin C) or the presence of covalently bound topoisomerases (fluoroquinolones) (24). The situation with ribosome inhibitors is more complex, because some such as kanamycin and streptomycin act as competence inducers (Fig. 2, B and C), whereas others such as erythromycin and tetracycline do not (fig. S3, A and B). A parallel can be made with the situation in E. coli: Kanamycin and streptomycin, which leave the ribosomal A site empty, induce a heat-shock–like response, whereas erythromycin and tetracycline, which either fill the A site with aminoacyl–transfer RNA (tRNA) or block it, trigger a cold-shock–like response (23). The level of ppGpp, which is produced by ribosomes that are stalled by a lack of charged tRNAs, decreases upon the addition of erythromycin or tetracycline (or a reduction in growth temperature) and increases as the growth temperature is increased (23, 25), suggesting that this nucleotide could act as a signal (4).

Whatever the underlying mechanism(s) may be, the induction of the com regulon by various antibiotics supports the hypothesis that competence is a general stress response of S. pneumoniae (4, 7) and is consistent with the proposal that CSP is not an effector of quorum sensing (7) but is an alarmone that conveys a stress signal (4). As a coordinator of competence, CSP enhances the efficiency of transformation as a rescue process in two ways: by increasing the number of potential transformants (3) and by triggering fratricide (5) through allolysis (defined as lysis in trans) and the release of DNA from CSP-nonresponsive pneumococcal cells. Whenever these cells differ from the competent population (for example, during cocolonization), allolysis provides a source of genetically diverse DNA. Through the coupling of fratricide and transformation, CSP could thus play a crucial role in generating genetic diversity under stress conditions for a species that seems unable to rely on inducible mutagenic repair (such as the SOS response) (7). Consequently, the high incidence of asymptomatic carriage of this pathogen is a major concern, because inappropriate antibiotic treatments could accelerate the occurrence of additional resistant clones and promote the evolution of virulence.

Supporting Online Material

Materials and Methods

Figs. S1 to S4

Tables S1 and S2


References and Notes

View Abstract

Navigate This Article