|Home | About | Journals | Submit | Contact Us | Français|
The ParA and ParB family proteins are required for accurate partitioning of replicated chromosomes. The Mycobacterium tuberculosis genome contains parB, parA and two parA homologs, Rv1708 and Rv3213c. It is unknown if parA and its homologs are functionally related. To understand the roles of ParA and ParB proteins in M. tuberculosis cell cycle, we have evaluated the consequences of their overproduction and visualized their localization patterns in M. smegmatis. We show that cells overproducing of ParA, Rv1708 and Rv3213c and ParB are filamentous and multinucleoidal indicating defects in cell cycle progression. Visualization of green-fluorescent protein fusions of ParA and its homologues showed similar localization patterns with foci at poles, quarter-cell, midcell positions and spiral-like structures indicating that they are functionally related. On the other hand, the ParBGFP fusion protein localized only to the cell poles. The cyan and yellow fluorescent fusion proteins of ParA and ParB, respectively, colocalized at the cell poles indicating that these proteins interact and possibly associate with the chromosomal origin of replication. Collectively our results suggest that the M. tuberculosis Par proteins play important roles in cell cycle progression.
Mycobacterium tuberculosis, the causative agent of tuberculosis, is responsible for 1.7 million deaths annually (WHO report 2008: Global tuberculosis control). The existence of multi-drug-resistant TB strains and the recent emergence of an extensively drug-resistant strain emphasize the need for novel drug therapies. The bacterial cell cycle is a complex process that includes chromosomal DNA replication, segregation of the newly replicated chromosomes into daughter cell compartments and cytokinesis. The division site is selected between the segregated chromosomes and the cell septum forms 1. Disruption of the chromosome segregation process can in turn disrupt cell division, causing aberrant septum formation and therefore, generation of anucleate cells 2. Latency, a hallmark of TB, describes a persistent state, in which M. tuberculosis possibly divide slowly but are restrained by the host immune system 3. As M. tuberculosis can shift between active and restricted growth in vivo, the knowledge of chromosomal segregation with respect to the cell cycle may help to elucidate a mechanism for this transition.
Plasmid and chromosomal partitioning are mediated by orthologs of ParA and ParB proteins 4-8. The parA gene encodes a dynamic protein that exhibits ATPase activity and the parB gene encodes a protein that binds DNA and forms a nucleoprotein complex at specific regions, called parS sequences, present on the plasmid or the chromosome near the origin of replication 2. M. tuberculosis genome encodes parA (Rv3918), parB (Rv3917), and two parA-like homologs, Rv1708 and Rv3213 (http://genolist.pasteur.fr/TubercuList). Himar-1 transposon mutageneis indicates that both parA and parB are essential for the growth of M. tuberculosis 9. Interestingly, the M. smegmatis parB counterpart is not essential and its overproduction has no phenotype 10. ParB protein has been shown to bind to three origin-proximal parS sequences and in vitro this binding is enhanced by interaction with ParA 10.
It is unknown if the Rv1708 and Rv3213 proteins are functionally related to ParA and how ParA, its homologs and ParB (collectively referred to as Par proteins) localize in actively growing cells. In this study, we have characterized the M. tuberculosis Par proteins with respect to their cellular localization and overexpression in M. tuberculosis and M. smegmatis. Our data indicate that M. tuberculosis ParA and its homologs show similar localization patterns and that ParA and ParB colocalize at cell poles suggesting that they interact in the partitioning complex in vivo.
M. smegmatis and M. tuberculosis strains were grown in Middlebrook 7H9 broth (Becton-Dickenson) supplemented with oleic acid-albumin-dextrose-catalase (OADC) (Becton-Dickenson) and Tween 80 (Sigma). Kanamycin at 25 μg ml-1 or hygromycin at 50 μg ml-1 were added as needed. All plasmids were propagated in E. coli as described11. Viability was measured by spreading appropriate dilutions of cultures grown with and without 0.2% acetamide for 6 hours on 7H10 plates with appropriate antibiotics and determining colony-forming units (CFU) as described 11.
parA, parB, Rv1708 and Rv3213 genes were amplified by polymerase chain-reaction from M. tuberculosis genomic DNA using gene-specific primers with appropriate restriction sites (Table 1) and cloned under the inducible acetamidase promoter in either pJAM2, a replicating shuttle vector, or pJFR19, an integrating vector 11. Expression was induced by the addition of acetamide to 0.2%. As needed, gfp, cfp or yfp was cloned into constructs containing the par genes (Table 2). The recombinant plasmids were electroporated into M. smegmatis and M. tuberculosis and transformants were selected on 7H10 agar plates containing appropriate antibiotics. To construct a strain producing ParA-CFP and ParB-YFP, we transformed parA:cfp into M. smegmatis bearing parB:yfp (Table 2).
Protein sequences for M. tuberculosis ParA (Rv3918), ParB (Rv3917), Rv1708 and Rv3213c (http://genolist.pasteur.fr/TubercuList/) were aligned using the ClustalW2 alignment tool (http://www.ebi.ac.uk/Tools/clustalw2/index.html).
Nucleoids were stained with a mixture of ethidium bromide (40 μg/ml) and mithramycin A (180 μg/ml) as described 12. For some experiments, actively growing M. smegmatis cells were incubated with the membrane dye, FM4-64 at a final concentration of 0.5 μg/ml for 1 to 1.5 hours. Next, cells were collected and stained with DAPI (Invitrogen) at 10 μg/ml for 5 to 10 min. The cells were visualized by brightfield and fluorescence microscopy using a Nikon Eclipse E600 microscope with a CoolSnap ES CCD camera (Photometrics) and a high-pressure mercury lamp (Nikon). A Nikon G-2E/C TRITC filter set (Ex528-553/Em600-660) was used for ethidium bromide/mithramycin A and FM4-64 –stained cells and a Nikon B2A filter set (Ex450-490/Em515) was used for GFP strains. The CFP and YFP strains were imaged using a Nikon CFP filter (Ex490-510/Em520-550) and Nikon YFP filter (Ex426-446/Em460-500). DAPI-stained cells were imaged with a standard DAPI filter set (Ex325-375/Em435-485, Chroma Technology). Images were analyzed using MetaMorph 6.2 software (Universal Imaging Corporation). At least 100 cells from each set were scored for staining patterns and cell length measurements.
Multiple alignment of the M. tuberculois ParA with Rv1708 and Rv3213c revealed high similarity among the three proteins. All three also have the conserved Walker A and B motifs, characteristic of several proteins that bind and hydrolyze ATP 13 (Fig. 1). The M. tuberculosis and M. smegamtis ParA proteins show significant sequence similarity (not shown).
The recombinant M. smegmatis strains overexpressing parA, parB, Rv1708 and Rv3213 were stained with DAPI and visualized by fluorescence and brightfield microscopy (Fig. 2). In wild-type M. smegmatis MC2155, ~ 90% of cells had 1 to 4 nucleoids (as shown in Figs. 2 panels 1c and 5b) and only 10% showed >4 nucleoids/cell. The cells overproducing ParA were elongated and multinucleoidal with more than 4 nucleoids per filamentous cell (Fig. 2, panels 2c and 3c). Similar results were observed with Rv1708 and Rv3213c overproducing cells (data not shown). ParB overproduction resulted in 54% of nucleated cells with 5 to 8 nucleoids and 17% with more than 8 nucleoids/cell. However, in both cases nucleoids appeared more diffuse as compared to wild type suggesting defects in DNA segregation. This may, in turn, give appearance of multinucleoidal phenotype. About 1 to 3 % wild-type and the recombinant strains overproducing ParA and its homologous were anucleate. In contrast, 26% of ParB overproducing cells were short, i.e. ~ 2μm long and anucleate (Fig. 2 panels 6a and 6b). These latter results are similar to those reported for P. aeruginosa cells overproducing ParB 14 and B. subtilis spo0J null mutants 6.
Visualization of septa following staining with FM4-64 and nucleoids by DAPI revealed that the wildtype cells contained midcell septa with well-segregated nucleoids in both halves of the cell (Fig. 2 panels 1a-d). Some of the Par overproducing cells contained midcell septa with distinct nucleoids (not shown) while others were without distinct or developing septa at midcell position (Fig. 2, compare panels 1b with 2b, 3b and 4b). These results indicate that the cell division process is delayed under Par overproduction conditions. We also noted that approximately 5% ParB (Fig. 2, panel 4b), 1% ParA (Fig. 2, panel 2b) and 3. 5% Rv1708 and Rv3123c (data not shown) cells contained septa at quarter-cell position (Fig. 2, see panel 2b, 4b). Thus, asymmetric septa presumably did not contribute much to the number of anucleate cells in cells overproducing ParB (Fig. 2-2a-d).
Our results showing accumulation of anucleate cells upon ParB overproduction suggest that these cells show reduced viability. Determination of viability by plating for CFU revealed that ParB overproduction led to a ~50% decrease in viability whereas that of the ParA, Rv3213c, or Rv1708 overproduction had little effect as compared to wild-type M. smegmatis (Fig. 3). Elevated levels of ParB cause growth inhibition in P. aeruginosa 14. Presumably, an optimal ratio of ParB to ParA is required for optimal cell cycle progression in M. smegmatis, much like the scenario observed with the B. subtilis Soj (parA family) and Spo0J (ParB family) proteins 15. M. smegmatis ΔparB strain shows delayed growth and a ~10-fold increase in the number of anucleate cells 10.
To further examine the roles of M. tuberculosis ParA and ParB in DNA segregation, we tagged ParA (and its homologs) and ParB to GFP and visualized their cellular localizations in M. tuberculosis and M. smegmatis by fluorescence microscopy (Figs. 4-I and 4-II, respectively). As the various types of localizations in the two strains were similar, further observations were restricted to M. smegmatis. M. tuberculosis ParA:GFP and the ParA homolog-GFP fusions formed several similar types of localizations; these included polar foci (uni or bi -polar, Fig. 4-IIA’), foci at the quarter-cell (Fig. 4-IIB’) and midcell positions (Fig. 4-IIC’) and a spiral-like distribution (Fig. 4-IIE’). Presumably, ParA, like Soj protein of B. subtilis, is capable of dynamic pole-to-pole and internucleoid movement 16. The Rv1708:GFP fusion localizations were similar to ParA, however, the spiral-like structures were observed more frequently (37% of cells, not shown). Finally, the Rv3213c:GFP showed a unique membrane localization (Fig. 4-IIF’). Since Rv3213c is predicted to be soluble, the membrane localization could be due to interactions with another membrane-bound protein. M. tuberculosis ParB:GFP primarily formed foci at one or both cell poles and only occasionally localized to the midcell (Fig. 4-II G’,H’, and K’, respectively). Polar localization indicates possible association with the chromosome origin, which is oriented toward the cell poles during cell division.
To observe the localization of the Par proteins with respect to nucleoids, cells were stained with a mixture of ethidium bromide/mithramycin to visualize nucleoids (Fig. 5-I). ParA-GFP was found to be dispersed between or occasionally overlapping the condensed nucleoids (Fig. 5-Ia-d), which suggests possible colocalization with the chromosome. Similar results were noted with Rv1708-GFP and Rv3213c-GFP (not shown). B. subtilis Soj forms foci associated with the nucleoids and exhibits dynamic assembly and disassembly of these foci 16. M. tuberculosis ParB was observed as distinct spots at the cell poles; however, a distinct association with condensed nucleoids was not apparent (Fig. 5-Ie-h). B. subtilis Spo0J is known to specifically associate with the region of the nucleoid proximal to the cell pole during both vegetative growth and sporulation 17, 18.
In order to explore the possibility that M. tuberculosis ParA and ParB colocalize, we created M. smegmatis strains producing ParA:CFP and ParB:YFP protein individually and together and visualized fluorescent fusion proteins. We routinely observed less ParA localization in the double strain when compared to strains in which only ParA-CFP was being overproduced (data not shown). Only 5% of cells expressed both the fusion proteins and ParA-CFP and ParB-YFP were colocalized at the cell pole in 74% of these cells (Figure 5-II). The double strain also grew much slower and exhibited extensive clumping in broth (not shown). Since ParB overproduction led to asymmetric septa formation, and since interactions between ParA and ParB are needed for segregation 10, 14, 19, our data indicate that the overproduction of both proteins presumably results in an enhanced interference with segregation, cell division and cell viability. Thus, our data also suggest that an optimal ratio of the two Par proteins is important for proper chromosome partitioning.
Collectively, our results suggest that M. tuberculosis ParA is a dynamic protein that possibly oscillates within the cell in a spiral-like formation and appears to colocalize with nucleoids and with ParB at one or both cell poles. Recently, Soj has been shown to dimerize and bind DNA in the presence of ATP 20. M. tuberculosis ParA appears to lack one of the two arginines implicated in dimerization and DNA binding and therefore these properties remain to be tested. Considering the localization patterns, it is tempting to speculate that Rv1708 and Rv3213 are capable of similar dynamic oscillation and association with the nucleoids, cell poles or ParB.
This work is supported by RO1-AI48417, RO1-AI41406 and R56-AI073966.