|Home | About | Journals | Submit | Contact Us | Français|
Condensins play a key role in global chromosome packing. Pseudomonas aeruginosa encodes two condensins, SMC-ScpAB and MksBEF. We report here that the two proteins are involved in the differentiation of the bacterium and impose opposite physiological states. The inactivation of SMC induced a state characterized by increased adhesion to surfaces as well as defects in competitive growth and colony formation. In contrast, MksB-deficient cells were impaired in biofilm formation with no obvious defects during planktonic growth. The phenotype of the double mutant was dominated by the absence of MksB, indicating that the observed growth defects are regulatory in their nature rather than structural. ATPase mutations recapitulated many of the phenotypes of the condensins, indicating their requirement for a functional protein. Additionally, inactivation of condensins dramatically reduced the virulence of the bacterium in a murine model of lung infection. These data demonstrate that condensins are involved in the differentiation of P. aeruginosa and reveal their importance for pathogenicity.
IMPORTANCE Adaptation and differentiation play key roles in bacterial pathogenicity. In Pseudomonas aeruginosa, an opportunistic human pathogen, these processes are mediated by the activity of an intricate regulatory network. We describe here novel members of this network, condensins. We show that the two P. aeruginosa condensins specialize in the establishment of the sessile and planktonic states of the bacterium. Whereas condensins have well-established roles in global chromosome organization, their roles in regulating bacterial physiology have remained unknown. Our data indicate that the two programs may be linked. We further show that condensins are essential for the pathogenicity of P. aeruginosa.
Pseudomonas aeruginosa is an opportunistic human pathogen that presents serious problems to patients with impaired immunity, wound infections, secondary lung infections, or who are in intensive care units (1, 2). Its virulence is largely attributed to its high intrinsic antibiotic resistance and adaptability, which help it occupy diverse niches. Strikingly, the bacterium can undergo clonal diversification and produce colonies with distinct morphologies, such as small-colony variants (SCVs) and biofilms. In particular, SCVs are often found in clinical isolates of P. aeruginosa and can be collected off biofilms or emerge in vitro following antibiotic treatment. They are characterized by the formation of small colonies with clearly defined edges, delays in planktonic growth, increased clumpiness, and sometimes increased antibiotic resistance (3,–5). SCVs retain their phenotypic traits over many generations even in the absence of the initial signal but—at least in some cases—eventually revert to their dominant cell morphology (6). These features bear resemblance to bacterial differentiation. The mechanism that leads to clonal diversification and the subsequent stabilization of the resulting cell lines appears to involve the activity of the signaling network (7), transcriptional positive feedback loops (8), and mutations (3, 9) but remains poorly understood on a causative level. Here, we describe a novel factor that affects P. aeruginosa differentiation, its condensins.
Condensins play a unique role in global chromosome organization (reviewed in references 10, 11, and 12). These multisubunit ABC-type cellular ATPases (Fig. 1A) act as macromolecular clamps that bridge distant DNA segments (13) and have an intrinsic ability to self-organize into chromosome scaffolds (14). Condensins of the structural chromosome maintenance (SMC) superfamily have been implicated in segregation of the origins of DNA replication (15,–17). The DNA reshaping activities of condensins reside in their core SMC subunits, whereas non-SMC subunits are critical for the regulation and subcellular recruitment of the proteins (18,–20). In several bacteria, inactivation of condensins leads to dramatic chromosome segregation defects and a reduction in cell viability (21,–23).
ATP modulates the activity of the complex by altering its architecture and interaction with DNA (18, 24,–26). Several distinct intermediates of the ATPase cycle can be trapped by specific mutations. The Walker B EQ mutation stabilizes the dimeric form of the SMC head domain (18, 26). The C motif SR mutation traps the ATP-bound monomeric form of the head (26, 27). The Walker B DA mutant and the Walker A KI and KD mutants are deficient in binding to ATP or DNA (13, 27).
P. aeruginosa encodes two condensins from two distinct superfamilies, SMC-ScpAB and MksBEF (28). MksBEF (PA4684 to PA4686) is encoded as a single operon (Fig. 1B). SMC (PA1527) is encoded downstream of a GntR family transcriptional regulator PA1526 and is separate from ScpA (PA3198) and ScpB (PA3197). Mutational inactivation of SMC results in low-frequency anucleate cell formation and chromosome packing defects (28, 29). We report here that both condensins are expendable under laboratory conditions in P. aeruginosa but are needed for the differentiation of the bacterium. MksB-deficient cells were impaired in biofilm formation, whereas Δsmc cells displayed defects during planktonic growth and bore many traits typical for other small-colony variants. Notably, inactivation of MksB suppressed the growth defects of Δsmc cells, indicating that the observed phenotypes are related to regulation rather than structural deficiencies. Finally, we report a dramatically reduced virulence of condensin-deficient P. aeruginosa in a mouse model of lung infection.
All experiments were done in accordance with the protocols approved by the Oklahoma State University Institutional Animal Care and Use Committee (VM-14-1). These protocols are in compliance with the U.S. Animal Welfare Act and the guidelines of the National Research Council.
P. aeruginosa strains used in this study are summarized in Table 1. PAO1-Lac (ATCC 47085) was used as the wild-type (WT) strain unless otherwise noted. Mutations were introduced using allele replacement as described earlier for OP103 (28), followed by the excision, when indicated, of the gentamicin resistance marker (Gm) with the help of the pFLP2 plasmid (30). Deletions of sspB were constructed using the pEXG2-based plasmids as described elsewhere (31). All strains were verified by PCR to confirm correct replacement.
OP107 (Δsmc) was constructed by the excision of the gentamicin resistance gene from OP103. OP108 and OP109 contain deletions of mksB (PA4686) between nucleotides 33 and 2331. To construct OP113 (Δsmc::ΔGm ΔmksB::ΔGm), the Δsmc::Gm trait was transferred into OP109 and was followed by the removal of Gm. OP110 and OP114 each encode an isopropyl-β-d-thiogalactopyranoside (IPTG)-inducible SMC (PA1527). In these strains, 18 nucleotides upstream of PA1527 (7 nucleotides downstream from the stop codon of PA1526) were replaced with the lacIq-PT7 cassette (32). OP115 harbors the lacIq-PT7 cassette upstream of PA1526. In OP111, lacIq-PT7-smc was introduced into OP109. OP116 strains harbor point mutations in the ATPase site of the endogenous smc as indicated. Tested point mutations were K37D (KD), D1092A (DA), E1093Q (EQ), and S1064R (SR). OP130 was constructed using plasmid pEXG2-SspB, which was a gift from S. L. Dove (31). In OP131 and OP132, the endogenous mksB was replaced with its DAS4-tagged version (31). The ATPase mutations of MksB (strains OP117) included D864A (DA), E865Q (EQ), and S829R (SR). These mutations were constructed using overlap extension PCR followed by the insertion of the fragment into plasmid pEX_ΔmksB (28) between the KpnI and BlpI sites. This was followed by the integration of the resulting mutants onto the endogenous mksB locus of the chromosome of PAO1.
pNPA_SMC is a pBAD-based plasmid that encodes SMC-HisGlyHis8. pNPA_SMC* carries point mutations in the ATPase site of SMC, which were introduced using the QuikChange site-directed mutagenesis kit, and was used for purification of mutant SMC. pUCP-SspB encodes arabinose-inducible SspB, which was derived from pUCP-MksB (28) by replacing mksB with sspB.
Bacteria were grown in LB or M9 medium supplemented with 0.4% Casamino Acids (M9CA) and 0.4% glycerol as indicated. Antibiotics were used at the following concentrations: 30 μg/ml gentamicin and 200 μg/ml carbenicillin for P. aeruginosa and 10 μg/ml gentamicin and 100 μg/ml ampicillin for Escherichia coli.
To evaluate competitive growth, overnight cultures of the indicated gentamicin-resistant mutant and gentamicin-sensitive PAO1 cells were mixed in equal ratios, where 2 × 104 cells were inoculated into fresh LB medium and further incubated at 37°C. The cell mixture was then reinoculated every 8 h (optical density at 600 nm [OD600] of 0.5). Aliquots of the mixture were removed at the indicated times and were plated on LB and LB plus gentamicin plates to quantify the mutant and total cell numbers. The strain loss rate was determined by fitting the data to single exponential decay.
Biofilm formation on polyvinyl chloride (PVC) was evaluated as previously described (33). Stationary cells were diluted 1:100 in M9CA medium containing 0.4% glycerol, dispensed at 100 μl into each microplate well, and incubated at 37°C for the indicated times. The cells were stained with 20 μl of 0.1% crystal violet for 10 min and then removed. The wells were washed three times with phosphate-buffered saline (PBS) and air dried for 15 min. The remaining dye was resuspended in 100 μl of 95% ethanol and quantified by measuring the absorbance at 600 nm. Cell adhesion to PVC was quantified in essentially the same way, except that the dispensed cells were at an OD600 of 0.6 and their incubation was for 5 min or 1 h as indicated in the text.
Exopolysaccharide production was measured using the Congo red assay (34) with minor modifications. Briefly, cells were washed and suspended in PBS at an OD600 of 3, supplemented with Congo red at 0.002% (mass/vol) for exponential cells or 0.005% for stationary cells, incubated for 30 min, and then pelleted by centrifugation. The amount of the remaining Congo red was determined by measuring the OD490 of the supernatant. These values were then normalized by comparing the absorbance values to those of a serially diluted Congo red stock.
For controlled degradation of MksB, we employed the ClpXP-dependent degron system (31). OP132 cells encode only the DAS4-tagged version of MksB and lack the endogenous adaptor protein SspB, a protein required for ClpXP-dependent proteolysis (35) (Table 1). Transformation of OP132 cells with the pUCP-SspB plasmid (degron on) but not the vector alone (degron off) resulted in nearly complete degradation of MksB-DAS4. Due to the leaky expression of SspB, MksB-DAS4 degradation was nearly complete in the presence and absence of the arabinose inducer. Similar results were obtained using a pPSPK-SspB plasmid (31), which encodes an IPTG-controlled SspB (data not shown).
MksB and MksEF were purified as previously described (28). SMC-His8 was cloned into the pBADN plasmid, overproduced. and purified using nickel chelate and heparin chromatography as described earlier for MksB (28). The main modification of this procedure was the use of Tris-Cl, pH 8.5, as a buffer at all steps and an additional gel filtration step on a Sephacryl S-300 column. The protein was dialyzed against 20 mM Tris-Cl (pH 8.5), 300 mM NaCl, 2 mM EDTA, 1 mM dithiothreitol (DTT), and 50% glycerol for storage. Rabbit polyclonal antibodies against SMC, MksB, and MksE were raised at Covance and affinity purified as described previously (28). ATPase rates were measured using the Molecular Probes EnzChek phosphate assay kit in 20 mM HEPES (pH 7.7), 40 mM NaCl, 2 mM MgCl2, 1 mM DTT, 5% glycerol, and 1 mM MgATP.
The mouse model of P. aeruginosa lung infection was adapted from previously published protocols (37). Bacteria were grown at room temperature up to an OD600 of 0.5 and were then collected by centrifugation, washed once in PBS, and resuspended again in PBS. Anesthetized (isoflurane) 10- to 12-week-old female mice of the C57BL/6 strain (Charles River) were intranasally inoculated with 3 × 106 CFU of wild-type or mutant P. aeruginosa and observed at least twice daily. One mouse in each group was intranasally inoculated with PBS only and treated as a control. No signs of infection were detected in the controls. Euthanized mice were immediately necropsied, and lung tissue was collected for histology and bacterial load. For histological analysis, 2-mm sections of lungs were fixed in formalin, paraffin-embedded, sectioned, and stained with hematoxylin-eosin. The tissues were microscopically evaluated by a board-certified veterinary pathologist who was blinded to the experimental groups. The remainder of the lung tissue was homogenized in PBS and plated on LB to determine the colony count.
To delineate the roles of condensins in the physiology of P. aeruginosa, we generated in-frame deletions of SMC and MksB (see Materials and Methods). Surprisingly, these deletions failed to produce dramatic defects in bacterial viability. When tested for colony formation, ΔmksB and Δsmc ΔmksB cells were virtually identical to the parental cells. The Δsmc strain, however, deviated by producing smaller and fewer colonies (Fig. 1C), which differed from the other strains in their structure. They had well-defined edges devoid of the large translucent zone typical for the wild-type P. aeruginosa (Fig. 1D). This appearance is similar to other small-colony variants of P. aeruginosa obtained by other means (4,–6). In contrast, ΔmksB cells appeared to be identical to the wild type in colony formation as did ΔmksB Δsmc cells. The latter result is notable and indicates that SMC and MksB play opposite roles in regulating P. aeruginosa physiology.
The frequency of colony formation by Δsmc cells was reduced 2.6-fold compared to the parental strain (Fig. 1E). Curiously, it was observed at 23°C but not at 37°C. This reduction could be complemented by constitutive expression of the smc gene from a pUCP22 plasmid but not the plasmid alone. When visualized by microscopy, Δsmc cells showed a propensity to form short chains of up to four cells at 23°C but not at 37°C (Fig. 1F). Thus, the reduction in colony formation was likely caused by a delay in the separation of cells following division.
Immunoblot analysis revealed that SMC and MksBEF are constitutively expressed in growing and stationary-phase cells at approximately 100 copies of SMC and MksE and 3,000 copies of MksB (Fig. 2A and andB).B). Although somewhat larger, the excess of MksB over MksE is not unlike that observed for the E. coli MukBEF during stationary phase and is consistent with the dynamic organization of the complex (19). The copy numbers of the condensins were not affected by the deletion of their counterparts (Fig. 2C), indicating that the proteins do not directly affect each other's expression.
Mutations in smc and mksB had different effects on bacterial fitness. Δsmc cells but not ΔmksB cells performed poorly during competitive growth (Fig. 3A). When cultured together with the parental strain, Δsmc cells were lost within several reinoculations. No such loss occurred when the mutants were cultured alone. The strain loss rate was markedly reduced when reinoculations were carried out with reduced frequency (Fig. 3C) and was completely absent during the stationary phase. Therefore, the reduced competitiveness of the mutant develops during cell growth.
Competitive growth defects were partially complemented by the inducible expression of SMC. To this end, we examined several constructs. First, we transformed the Δsmc cells with a pUCP22-based plasmid pUCP_SMC, which encodes SMC under the control of an IPTG-inducible PT7 promoter at approximately 4× its endogenous level (Fig. 3B). A significant reduction in the strain loss rate was observed with the addition of a plasmid harboring the smc gene; however, no rate change occurred due to the vector alone (Fig. 3C). We next introduced a PT7 promoter into the intergenic region between PA1526 and SMC. Depending on the amount of IPTG, the resulting cells produced SMC below or above its endogenous level (Fig. 3B). In all cases, we observed partial complementation of the competitive growth defects in the presence, but not in the absence, of IPTG whether or not mksB was present inside the cell (Fig. 3C). Thus, the residual fitness defects in Δsmc cells were likely due to some cis consequences of the deletion rather than the amount of the expressed protein. Accordingly, no loss of fitness was detected in OP115 (lacIq-PT7-PA1526) cells, which produce SMC from the PT7 promoter placed upstream to PA1526 (Fig. 3C). Under noninducing conditions, however, these cells produced notably lower competitive growth defects, which is perhaps due to an interaction between SMC and PA1526 or the leaky expression of smc.
Only SMC-deficient cells, not other condensin mutants, displayed anomalies during planktonic growth. They remained clumpy for several hours after transfer into fresh medium, resulting in a characteristic lag in the growth curves (Fig. 3D). This lag may potentially be the reason why the strain performed poorly during competitive growth.
Although fully functional during competitive growth, MksB-deficient cells displayed a markedly reduced ability to form biofilms both on glass and on polyvinyl chloride (PVC) surfaces (Fig. 4A). In contrast, Δsmc cells were, if anything, more proficient in producing biofilms than the parental strain (Fig. 4B). This was mostly due to the high adherence of Δsmc mutants to the surface. Indeed, even 5 min of incubation of Δsmc but not PAO1 cells resulted in significant cell attachment to the PVC surface (Fig. 4C). Similar to the case with colony formation (Fig. 1), the phenotype of the double mutant was dominated by the absence of mksB.
Depletion of MksB using the degron system (31) also resulted in impaired biofilm formation, demonstrating that this effect was indeed due to the absence of the protein (Fig. 4D and andE).E). To this end, we removed sspB, which encodes the protease adapter protein, from PAO1 and replaced the endogenous mksB with another version that encodes a C-terminal DAS4 tag (31). The resulting strain produced a tagged MksB at its endogenous level until it was transformed with a plasmid that harbors sspB (Fig. 4D). As expected, biofilm formation was impaired in all strains where MksB production was depleted (Fig. 4E). Because the cell physiology changes in response to the production of the plasmid-borne SspB protein, we conclude that the phenotype is associated with the presence of MksB.
The stickiness of P. aeruginosa can often be attributed to the production of exopolysaccharides by the Pel/Psl system (38). To determine whether or not the same holds true for our mutants, we employed the Congo red assay (34, 39). The assay exploits the high affinity of the dye to polysaccharides. The cells were supplemented with Congo red for 30 min and pelleted by centrifugation, and the remaining Congo red was spectrophotometrically quantified (Fig. 4F). We found that Δsmc cells produce Pel/Psl on par with the wild type in the stationary phase, but unlike the wild type, they fail to switch it off upon the resumption of growth (Fig. 4G). In contrast, ΔmksB and ΔmksB Δsmc cells downregulate the synthesis of polysaccharides both during exponential growth and in the stationary phase. Thus, the adhesion behaviors of the condensin mutants are mediated by the Pel/Psl system.
ATP is central to the operation of condensins. We generated point mutations in the ATPase site of SMC as a means of modulating its activity and exploring correlations between its phenotypes. We tested the following three types of mutations (Fig. 5A): (i) the Walker B DA and Walker A KD mutations, which preclude ATP binding and dimerization of the SMC heads (13, 27); (ii) the Walker B EQ mutation, which precludes ATP hydrolysis but not binding and stabilizes the dimeric SMC heads (18, 26); and (iii) the C motif SR mutation, which interferes with head dimerization but not ATP binding (26, 27). All tested mutants were expressed in P. aeruginosa at their normal level (Fig. 5B) and, for the case of SMC, confirmed for their deficiency in ATP hydrolysis (Fig. 5C).
The Walker B mutant SMCEQ was indistinguishable from the wild-type protein in all tested assays, including anucleate cell formation (Fig. 5D), colony formation (Fig. 5E), and competitive growth (Fig. 5F). All other mutants were deficient in one or another aspect of SMC activity. In particular, DA, SR, and KD mutations prompted the formation of anucleate cells at the level found in the deletion mutant (Fig. 5D). However, defects in colony formation and competitive growth were observed only for SMCKD (Fig. 5E and andF).F). All of these mutations interfere with the dimerization of the SMC heads, whereas the charge inversion KD mutant is deficient in DNA bridging (13, 26, 27).
These results reveal that the phenotypes of SMC mutants are only partially related. For example, the DA and SR mutants of SMC produced anucleate cells but did not show any fitness defects. It appears, therefore, that at least some of the regulatory functions of SMC are unrelated to their role in chromosome packing.
ATPase activity was similarly essential for the function of MksB. Two out of three tested ATPase mutants, EQ and SR, were deficient in biofilm formation (Fig. 5G). However, one of the variants, MksBDA, failed to produce the said phenotype. This is not unlike the EQ mutant of SMC, which behaved similarly to the wild type in all fitness assays (Fig. 5D to toF).F). It appears, therefore, that the regulatory functions of condensins can be performed by one of their intermediates in the ATPase cycle.
Even though the phenotype of condensin mutations appeared rather mild, their effect on epigenetic states was quite dramatic. To determine whether or not these changes alter the virulence of P. aeruginosa, we evaluated Δsmc and Δsmc ΔmksB strains in a mouse model of lung infection (37).
Eight C57BL/6 mice were each intranasally infected with 3 × 106 CFU of PAO1 (WT), Δsmc, or Δsmc ΔmksB cells, and their survival followed (Fig. 6A). All of the mice infected with PAO1 and Δsmc ΔmksB cells initially appeared clinically ill, which was primarily characterized by depression and increased respiratory rate. However, the Δsmc ΔmksB mice quickly stabilized, whereas the PAO1 mice continued to deteriorate and most had to be euthanized by day 1 postinfection. No clinical signs of illness were detected in the Δsmc mice. Thus, both condensin mutants were clinically less virulent than the parental strain based on survival data.
All PAO1 mice euthanized on day 1 had severe bacterial colonization of the lungs along with pulmonary hemorrhage and mild neutrophilic infiltrates (Fig. 6B). The inflammatory reaction appeared mild most likely because the mice were killed before significant infiltrates (neutrophils and macrophages) had time to accumulate. It seems like the lesions and death in the mice inoculated with the WT were more from hemodynamic events (hemorrhage, fluid leakage, and shock) related to the pathogen rather than any accompanying component of severe host inflammatory response.
The Δsmc mice all had pneumonias; however, their pneumonias were characterized by heavy cellular infiltrates (primarily neutrophils and macrophages) within the airways and alveolar spaces. The myriad bacterial colonies that were so conspicuous in the lungs of PAO1 mice were not seen in the Δsmc-infected animals (Fig. 6B). The Δsmc ΔmksB mice that lived until the end of the study had histological lesions that were indistinguishable from the single smc mutants, whereas the lone Δsmc ΔmksB mouse that had to be killed on day 2 had histological lesions that resembled the WT.
The bacterial load analysis was consistent with the histological results. The PAO1 mice had significantly higher bacterial loads than the Δsmc ΔmksB mice (1.8 × 107 versus 5.3 × 106 CFU/g), and the number of bacteria in the Δsmc lungs was below detection (less than 100 CFU/g) (Fig. 6C). According to the Student test analysis, the difference between the PAO1 and Δsmc ΔmksB mice is significant (P = 0.02) just as it is between the PAO1 and Δsmc mice (P < 2 × 10−4). The CFU count also formed distinct distributions for mice that fell to the sickness and those that survived it (P = 4 × 10−5), which is consistent with mortality being caused by infection.
Based on these results, we conclude that condensin-deficient P. aeruginosa are significantly less virulent than the parent. We further conclude that Δsmc P. aeruginosa is even less virulent than the double mutant Δsmc ΔmksB. This result, while somewhat counterintuitive, is fully consistent with the phenotypic profiles of the two strains, which revealed the enhanced fitness of Δsmc ΔmksB and ΔmksB cells.
P. aeruginosa is notorious for its ability to thrive in a hostile environment. To achieve this, it encodes numerous biosynthetic and regulatory enzymes that help it adapt to adverse conditions. Befittingly, P. aeruginosa harbors two condensins. One of them is a conventional SMC condensin while the other, MksBEF, belongs to a recently discovered and poorly characterized family that is distantly related to E. coli MukBEF. We show here that the two proteins are involved in the differentiation of the bacterium. This is a novel activity of condensins, which may or may not be related to their function in chromosome maintenance.
Inactivation of SMC and MksB had opposite effects on the physiology of P. aeruginosa. The ΔmksB cells were deficient in their adhesion to surfaces and the formation of biofilms (Fig. 4) but did not display defects during planktonic growth (Fig. 3). If anything, ΔmksB cells appeared to be better fit for planktonic growth since they reached high densities sooner than the parental cells (Fig. 3D). In contrast, Δsmc cells had marked fitness defects (Fig. 3) but were well adapted for sessile growth (Fig. 4). These effects appear to be regulatory rather than structural in nature since the phenotype of the double mutant was dominated by the absence of MksB. Additive effects would instead be expected if the inactivation of condensins resulted in the inactivation of an essential pathway. In this sense, MksB and SMC appear to be integrated into opposite developmental programs in P. aeruginosa (Fig. 7).
The phenotype of Δsmc cells was multifaceted and included such features as a decrease in competitive index (Fig. 3A), a lag in planktonic growth (Fig. 3D), increased adhesion to surfaces (Fig. 4B and andC),C), and the formation of small colonies with sharp edges (Fig. 1D). All of these phenotypes have been previously reported for other SCVs of P. aeruginosa, which were obtained by antibiotic treatment or isolated in clinics (3,–5). The commonality of these phenotypes, as well as the ability of SCVs and “normal” cells to slowly covert into each other, suggests that they represent a distinct differentiation state in P. aeruginosa, which can be induced by a variety of means. Curiously, many of these phenotypes can be attributed to a single trait, the expression of the exopolysaccharides by the Pel/Psl system (Fig. 4F and andG).G). Given the epigenetic nature of this transition and because it helps the bacterium to colonize new niches, it might be prudent to refer to this as differentiation. Our finding that this transition can be triggered by condensins, a low-copy-number motor protein, suggests new ways to induce differentiation programs.
Despite their relatively mild phenotype, condensin mutants had dramatically reduced virulence in a murine model of lung infection. Δsmc cells were completely cleared by the immune system without producing visible clinical signs (Fig. 6). Even the Δsmc ΔmksB mutant, which has fewer growth defects in planktonic bacteria, was much less virulent than the parental strain. These data demonstrate that condensins are essential for virulence in P. aeruginosa.
The finding that condensins participate in epigenetic control was unexpected. Indeed, these proteins contribute to an essential cellular function, chromosome replication and segregation. Their participation in an additional cellular program increases the costs of mitigating mutational pressure. Clearly, the benefits of maintaining such coordination must outweigh the costs. A plausible benefit from such coordination may involve adjusting chromosome structure to a particular differentiated form. Sporulating bacteria, for example, condense the chromosome in the spore but not the mother (40). Intriguingly, our data suggest that this link may also work in the opposite direction where information on chromosome packing can influence epigenetic switching.
We are indebted to Marie Montelongo and Pulavendron Sivasamiv for mouse inoculations and to Herbert Schweizer and Simon Dove for sharing strains and plasmids.
The content of this article does not necessarily reflect the position or the policy of the federal government, and no official endorsement should be inferred.