Distribution of the Escherichia coli structural maintenance of chromosomes (SMC)-like protein MukB in the cell



Fluorescent polyclonal antibodies specific for MukB have been used to study its localization in Escherichia coli. In wild-type cells, the MukB protein appeared as a limited number of oblong shapes embracing the nucleoid. MukB remained associated with the nucleoid in the absence of DNA replication. The centre of gravity of the dispersed MukB signal initially localized near mid-cell, but moved to approximately quarter positions well before the termination of DNA replication and its subsequent reinitiation. Because MukB had been reported to bind to FtsZ and to its eukaryotic homologue tubulin in vitro, cells were co-labelled with MukB- and FtsZ-specific fluorophores. No co-localization of MukB with polymerized FtsZ (the FtsZ ring) was observed at any time during the cell cycle. A possible role for MukB in preventing premature FtsZ polymerization and in DNA folding that might assist DNA segregation is discussed.


Structural maintenance of chromosomes proteins (SMC) are involved in the maintenance of higher order structures of chromatin. In eukaryotes, they are essential components of multiprotein complexes denoted as condensins and cohesins (Cobbe and Heck, 2000;Ball and Yokomori, 2001). They are ubiquitous in eukaryotes, although much less so in prokaryotes, which is probably related to the comparatively simple structure of the bacterial chromosome or nucleoid. The Escherichia coli 177 kDa MukB protein shows similarities in domain organization to eukaryotic SMCs (Niki et al., 1991) and to the Bacillus subtilis Smc protein (Melby et al., 1998). MukB forms an antiparallel coiled-coil homodimer (Melby et al., 1998). The N-terminal and C-terminal globular domains meet at each end of the coiled-coil region, which is intersected by a hinge region, and they share an ATP binding site. The N-terminal domain contains the conserved Walker A motif for NTP binding, and the C-terminal domain has been suggested to contain a Walker B motif (Niki et al., 1991). Based on electron micrographs, the MukB protein seems to exist in two conformations: a closed one and an open one (V-shaped) (Niki et al., 1992;Melby et al., 1998), suggesting that the molecule can make a scissoring movement. MukB forms a larger complex by binding MukF and MukE with its C-terminal domain (Yamazoe et al., 1999).

In vitro, MukB can bind DNA and ATP or GTP (Niki et al., 1992). Missense and null mukB mutants are temperature sensitive and produce anucleate cells with much higher frequency than wild-type cells (5–15% versus 0.03%) (Niki et al., 1991). The B. subtilis Smc protein has been shown to bind, aggregate and anneal single-stranded DNA in vitro (Hirano and Hirano, 1998). In vivo, it causes disruption of the nucleoid structure and dispersion of SpoOJ (Moriya et al., 1998). SpoOJ belongs to the ParB family, which is involved in plasmid and chromosome partitioning (Moriya et al., 1998).

Several observations suggest that MukB is involved in DNA condensation: (i) MukB mutants are repressed by topoisomerase I mutants that cause an increase in negative supercoiling resulting from gyrase A activity (Sawitzke and Austin, 2000); (ii) MukB mutants show dispersed elongated decondensed nucleoids; and (iii) mukB null mutations are suppressed by mutations in seqA, encoding a protein involved in the sequestering (decondensing) of newly replicated DNA (Lu et al., 1994; Weitao et al., 1999;Brendler et al., 2000). These observations suggest that, in order to carry out these activities, MukB would be closely associated with the nucleoid.

The N-terminal domain of MukB can bind to the bacterial tubulin homologue FtsZ (Lockhardt and Kendrick-Jones, 1998a), as well as to microtubules (Lockhardt and Kendrick-Jones, 1998b). It has been speculated that MukB could be a motor protein that ensures equipartitioning of the chromosomes into the new daughter cells using the FtsZ ring as a track (Hiraga, 1992; Lockhardt and Kendrick-Jones, 1998a). FtsZ initiates cell division by polymerizing into a ring-like structure at mid-cell (Bi and Lutkenhaus, 1991; Den Blaauwen et al., 1999). Because the mid-cell localization of all other known essential cell division proteins requires FtsZ (Weiss et al., 1999), the FtsZ ring was postulated to form the framework that recruits other components of the putative division machinery or divisome. Therefore, FtsZ is the most likely target in the cell for spatial and temporal regulation of this process.

After submission of this manuscript, a paper appeared also addressing the localization of MukB in E. coli (Ohsumi et al., 2001). These authors localized MukB by visualizing a MukB–GFPuv4 fusion protein or by immunolabelling the green fluorescent protein (GFP) part of MukB. They observed the fusion protein in the nucleoid as one or two ‘discrete fluorescent foci’, dependent on the cell cycle stage (Ohsumi et al., 2001). In contrast, we find by immunofluorescence microscopy of wild-type MukB a more dispersed distribution as oblong shapes in the nucleoid, which we believe to be more in line with its putative nucleoid-condensing function (Britton et al., 1998). MukB did not co-localize with the FtsZ ring, i.e. with polymerized FtsZ.


The mukB null mutant is not labelled

To localize MukB in the cell by immunofluorescence, polyclonal antibodies have been made against purified MukB. To test the cytological specificity of the antibodies, we compared the MukB labelling in wild-type cells and in a mukB null mutant. Because of the temperature sensitivity of the mukB null mutant, cells were grown in TY at 20°C with a mass doubling time (Td) of 100 min. A number of representative cells of wild type and null mutant are shown in Fig. 1. The top left part of this figure shows a clear MukB signal in the wild-type cells and no signal at all in the top right part of the figure with the mukB null mutant cells. Upon a 10-fold longer exposure, a weak signal could be observed in the null mutant that was similar to the signal after incubation with the secondary antibody only. This shows that the affinity-purified antiserum is highly specific for the MukB protein.

Figure 1.

Wild-type cells (MC4100) and MukB null mutant cells (GC7527 ΔmukB::Kan) grown in TY at 20°C. The cells were immunolabelled with MukB affinity-purified IgG R124519 against MukB and goat anti-rabbit Alexa 546 secondary antibodies. Left: MC4100 cells. Right: GC7527 cells. The top pictures show the localization of MukB using Alexa 546 fluorescence (illumination 0.8 s); the centre pictures show the DAPI-stained nucleoids (illumination was 0.2 s and 0.1 s for MC4100 and GC7527 respectively), and the bottom pictures show the phase-contrast image (illumination 0.2 s). No MukB signal could be detected in the mukB null mutant. Only after prolonged illumination (8 s) was a faint background observed in the cells, which was identical to that obtained using the secondary Alexa 546-conjugated antibody only. The colours are false. The bar is 1 µm.

MukB localization is restricted to the nucleoid region

As mentioned above, the cells shown in Fig. 1 had been grown in a rich medium (Td = 100 min at 20°C). In these cells, the MukB fluorescent signal appears to be restricted to the same region as the DAPI-stained nucleoids (Fig. 1, centre) and, within the limit of our detection, no free MukB molecules were seen in the cytoplasm. The fluorescent MukB signal is composed of oblong shapes, which often seem to be arranged into nucleoid-embracing configurations. Subsequent labelling studies were carried out with the MC4100 strain growing in steady state with a Td = 85 min at 30°C unless stated otherwise (for a definition of steady state, see Fishov et al., 1995).

The oblong MukB shapes often seemed to be quite extended while focusing in the axial direction. To substantiate this observation, 15 optical slices were made by confocal scanning laser microscopy of MukB-labelled MC4100 cells. The images have been corrected for the point spread function of the optics. Four optical slices moving from the upper left to the lower right image through the sample are shown in Fig. 2. The majority of the oblong structures seem to have a boomerang-like shape, with a length of about 700 nm. Whether these shapes are structural entities as such or whether they consist of multiple foci is hard to determine, because of the limiting resolution of the microscope.

Figure 2.

Images of optical sections of the MukB structures in MC4100 grown to steady state in GB1 at 30°C with a mass doubling time of 85 min. A stack of 25 optical sections of 0.1 µm through the bacteria was collected and corrected for the PSF of the optics (see Experimental procedures). Four optical sections from the upper left to the lower right show the MukB structures at 0.2 µm depth intervals. The arrows point at a MukB structure that consists of two connected foci, of which the left focus is moving out of focus, while the right one is moving in the optical section series. The bar represents 100 nm.

MukB in aztreonam-treated cells

To substantiate further the co-localization of MukB with the nucleoid, we compared their respective distribution in aztreonam-induced filaments. Aztreonam interacts specifically with PBP3 (Sykes and Bonner, 1985), thus inhibiting the completion of cell division. However, FtsZ rings can still form (Fig. 3B), and they even become stabilized in the presence of aztreonam (our unpublished observations). The replicating nucleoid pairs are evenly spaced along the aztreonam-induced filaments (Fig. 3A), and the MukB label (Fig. 3C) follows this pattern. Although MukB appears to be present in the nucleoid, its distribution within it is clearly not homogeneous (see Discussion).

Figure 3.

Aztreonam filaments of MC4100 grown to steady state in GB1 at 30°C with a mass doubling time of 85 min.

A. The nucleoids stained with DAPI.

B. The same cells labelled with Alexa 488-conjugated monoclonal antibody F168-12 Fabs against FtsZ.

C. The cells labelled with Alexa 546-conjugated polyclonal antibody Fabs R124519 against MukB.

D. A merger of the MukB and FtsZ signal.

The colours are false. The bar is 1 µm.

MukB localization during the cell cycle

We have followed two different approaches to measure the distribution of MukB during the cell cycle of wild-type MC4100. First, we determined the centre of gravity of the fluorescent intensity of MukB in a nucleoid and plotted the obtained values against cell length. In small cells, the centres of gravity tended to be in the cell centre, whereas at longer lengths, they distributed roughly near the one-quarter and three-quarter positions in the cell (Fig. 4A). In the population, 65% of the cells (n = 406) contained two centres of gravity. From the age distribution, it can be calculated (Den Blaauwen et al., 1999) that duplication of the centres of gravity takes place at a cell age of 21 min, corresponding to a cell length of about 1.9 µm. As determined before, newborn cells have replicated their DNA by 50% (Huls et al., 1999). Termination of DNA replication occurs ≈ 37 min after birth, which corresponds to a cell length of about 2.4 µm (Den Blaauwen et al., 1999). Thus, separation of the MukB centres of gravity takes place well before the termination of DNA replication. A disadvantage of this analysis is that the actual distribution of fluorescently labelled MukB within the nucleoid is obscured; the centre of gravity approach mimics aggregation of MukB.

Figure 4.

Distance of MukB centres of gravity from mid-cell in relation to cell length. Negative values refer to the arbitrary left side of a cell; positive values refer to the arbitrary right side of a cell. Grey lines represent the cell poles; dashed grey lines represent the one-quarter and three-quarter positions of the cells.

A. MukB centres of gravity in normally replicating MC4100 cells grown in GB1 medium at 28°C with a mass doubling time of 80 min, n = 406.

B. MC4100 cells growing in the presence of rifampicin for 90 min, n = 511. It should be noted that the rifampicin-treated cells are smaller because they have completed their division, and they are somewhat growth inhibited because of the absence of new mRNA synthesis.

In the second approach, we took into account the fluorescent intensity distribution within a cell (Fig. 5). In this presentation, the total cellular fluorescence of the DAPI-stained nucleoid and the immunostained MukB has been plotted on the length axis of cells selected according to size. In small cells, a single DAPI peak (blue curves) can be seen, which broadens when cells become longer. At a cell length of 2.7 µm or at 70 min into the cell cycle, nucleoids become physically separated. In contrast, separation of the fluorescent MukB distributions (green curves) occurs at a much smaller length (about 1.9 µm), which is in line with the centre of gravity approach (compare Figs 4 and 5).

Figure 5.

Correlation between the MukB signal and cell length in MC4100. Top to bottom in each section is the phase-contrast, DAPI fluorescence, MukB fluorescence, integrated density of the DAPI signal using a scanning rectangular window, integrated density of the MukB signal using the same scanning window and the bar representing the cell length in µm, with the corresponding cell age in min between brackets. The termination of DNA replication occurs on average at 37 min, and a new round of replication occurs on average at 52 min (Huls et al., 1999). The cells were grown to steady state in GB1 at 30°C with a mass doubling time of 85 min. The colours are false.

MukB remains associated with the nucleoid after inhibition of DNA replication

To answer the question whether the distribution of MukB within the nucleoid is related to DNA replication, we labelled MukB in wild-type MC4100 cells treated with rifampicin for 90 min. Under these conditions, there is run-out synthesis of DNA, and reinitiation of DNA replication becomes blocked (Bipatnath et al., 1998). With a DNA replication period of 70 min, not a single DNA replicating cell is expected to be present after 90 min of incubation with rifampicin (Huls et al., 1999). After this treatment, the nucleoids became somewhat rounded as seen before (Huls et al., 1999); however, the labelled MukB stayed with the DAPI-stained nucleoid, as judged from the centre of gravity approach (Fig. 4B). Here, a similar distribution of MukB can be seen as in replicating nucleoids (Fig. 4A).

In an additional experiment, we labelled E. coli B/rK cells that had been grown with a Td of 160 min at 37°C. Under these conditions, this strain has no DNA synthesis at the beginning of the cell cycle, i.e. during the so-called B-period (Koppes and Nanninga, 1980). Also in this case, fluorescently stained MukB occupied the same cellular area as the DAPI-stained nucleoid in small non-replicating cells (not shown). The two experiments thus indicate that there is no visible relation between the localization of MukB and the replication status of the DNA.

MukB does not co-localize with the FtsZ ring

The MukB localization pattern was not much different after incubation of the cells with the crude IgG compared with affinity-purified IgG. Presumably, the extra recognized protein band of 60 kDa on the immunoblot (see Experimental procedures) corresponds to a protein that is either not very abundant or is not recognized with high affinity in the native folded state. This enabled us to prepare Fab fragments of Protein A affinity-purified IgG (see Experimental procedures) conjugated with the Alexa 546 fluorophore, which allowed co-labelling of MukB and FtsZ.

Cells were labelled with polyclonal Fabs conjugated with Alexa 546 specific for MukB and monoclonal F168-12 Fabs conjugated with Alexa 488 specific for FtsZ. An example of the double-labelled cells is shown in Fig. 6, in which a cell with a deep constriction and one with a just initiated constriction are visible. A surface plot of the fluorescence intensity of the DAPI-stained nucleoids, the MukB fluorescence and the FtsZ fluorescence is shown to facilitate comparison of the fluorescence intensity distributions. The MukB fluorescence is maximal adjacent to the DAPI maximum, suggesting that the nucleoids are more or less embraced by two MukB structures. The FtsZ maximum does not show any overlap with the MukB signal. Out of 500 cells, no cell could be found in which overlap between the FtsZ ring and the MukB fluorescence could be seen. In aztreonam-induced filaments also, no co-localization of MukB and FtsZ rings was observed (Fig. 3D). Based on these observations, it seems unlikely that MukB binds to the FtsZ ring, i.e. to polymerized FtsZ.

Figure 6.

MC4100 cells labelled with Alexa 546-conjugated polyclonal antibody Fabs R124519 against MukB and with Alexa 488-conjugated monoclonal antibody F168-12 Fabs against FtsZ. The nucleoids were stained with DAPI. From top to bottom, the left pictures show phase-contrast image of two cells, the position of the nucleoids in blue, the MukB signal in green and the FtsZ rings in red. To the right, the corresponding fluorescence intensity surface plots are shown. The cells were grown to steady state in GB1 at 30°C with a mass doubling time of 85 min. The colours are false. The bar is 1 µm.


Multiple MukB molecules could form extended structures that embrace the nucleoid

In the wild-type strain MC4100 when grown in rich medium (Td = 100 min at 20°C; Fig. 1), in MC4100 cells (Td = 85 min at 30°C; Fig. 6) as well as in MC4100 cells treated with aztreonam (Fig. 3), the MukB localization pattern consists predominantly of oblong-shaped bodies within the confinement of the nucleoids. In the first case, the pattern appears more complex, presumably as a result of multifork replication. How many MukB molecules could be contained in an oblong-shaped body? The MukB protein itself has a length of up to 130 nm, dependent on the angle of the coiled-coils adjacent to the hinge (Melby et al., 1998). Complexed with MukF and MukE it has a Stokes radius of 12.3 nm (Yamazoe et al., 1999). Immunolabelling of 30–40 MukB molecules attached to a strand of DNA could easily result in an oblong structure of about 700 nm as observed. The total number of MukB molecules per cell has been estimated to be in the order of 150 (Kido et al., 1996). Although we do not know how this number compares with our growth conditions, it nevertheless indicates that enough MukB will be present to form roughly in the order of four elongated bodies during most of the cell cycle of MC4100.

Is MukB associated with the replisome?

Labelling experiments suggest that the replication machinery or replisome of the B. subtilis chromosome localizes to the cell centre (Lemon and Grossman, 1998; 2000). In the case of E. coli, it was found by electron microscopic autoradiography that [3H]-thymidine incorporation takes place in the cell centre, at least during the initial part of the DNA replication period (Koppes et al., 1999). This could mean that the replicating DNA strands are pulled through the replisome as was predicted by Dingman (1974). As a logical consequence of the replisome model, the terminus would be the last chromosomal region to end up at the cell centre, which has been observed (Gordon et al., 1997; Niki and Hiraga, 1998; Niki et al., 2000).

As has been deduced from the data in Fig. 4A, duplication of the centres of gravity of MukB fluorescence and their assumption of new cellular positions already occurs about 21 min after birth, thus before termination of DNA replication at about 37 min. This indicates that the bulk of MukB is not located near the terminus, and it argues against a replisome-associated function. A double labelling of MukB and the replisome would be required to elucidate this matter further. MukB remains associated with the nucleoid after run-out DNA replication (Fig. 4B) or when MukB is labelled during the B-period (see Results). This suggests, but does not prove, that MukB is not involved in DNA replication as such.

In a very recent study, the localization of SeqA has been compared with that of MukB. No co-localization was found during the major part of the cell cycle (Ohsumi et al., 2001). On the other hand, mutations in mukB and seqA suppressed each other's phenotypes (Weitao et al., 1999). The absence of co-localization of their gene products is in line with the earlier suggestion that suppression ‘occurs through one or more structure(s) or function(s) that are independently influenced by the mukB and seqA mutations’ (Weitao et al., 1999). So, it remains to be established whether specific DNA regions of the nucleoid are temporally or permanently associated with MukB.

Recent evidence shows that MukB (Sawitzke and Austin, 2000; Weitao et al., 2000a) and SeqA (Weitao et al., 2000b) affect chromosome superhelicity, although in an opposite way. In addition, the mukB phenotype could be suppressed by mutations in topA (Sawitzke and Austin, 2000). The main role of MukB might thus reside in assisting in correct folding of the nucleoid, thereby perhaps also acting as a pulling force to facilitate segregation (Sawitzke and Austin, 2000; Weitao et al., 2000a).

The localization of MukB and B. subtilis Smc is not the same

The B. subtilis MukB homologue Smc has been subcellularly localized by fluorescence microscopy as discrete nucleoid-associated foci that were ‘near a pole of the nucleoid’ in 70% of the cells using an Smc–GFP fusion (Britton et al., 1998). Using immunofluorescence, Smc was likewise found in the chromosome, but also at the cell poles, although not associated with the nucleoid (Graumann et al., 1998). The latter feature was particularly apparent in aseptate filaments, in which the Smc signal occurred in between the nucleoids (Graumann et al., 1998). This is quite different from our E. coli data, where we find MukB neither outside the nucleoid in normal cells (Figs 5 and 6) nor in filaments (Fig. 3).

MukB is not associated with the FtsZ ring

It has been reported that the N-terminal domain of MukB can bind to FtsZ (Lockhardt and Kendrick-Jones, 1998a) as well as to microtubules (Lockhardt and Kendrick-Jones, 1998b). Our data do not support this idea, at least with respect to polymerized FtsZ in the divisome. Normally, FtsZ ring formation does not occur in the presence of a centrally located nucleoid (nucleoid occlusion; Woldringh et al., 1990; Yu and Margolin, 1999). However, in a ΔmukB mutant, FtsZ rings can form on top of nucleoids (Sun et al., 1998). This was also observed in a dnaC2ΔmukB double mutant at the non-permissive temperature (Gullbrand and Nordström, 2000). It was also found that FtsZ forms aberrant structures in the cytoplasm, such as spirals and multiple rings, in a mukB null mutant (Sun et al., 1998). It is feasible that MukB interacts with cytosolic FtsZ, because it occupies a rather large area of the nucleoid. Perhaps MukB prevents premature polymerization of FtsZ at the site of division. We speculate that, in this way, MukB could contribute to the nucleoid occlusion effect.

Experimental procedures

Bacterial strains and growth conditions

Escherichia coli K-12 [MC4100 lysA (F, araD139, Δ(argF-lac)U169deoC1, flbB5301, ptsF25, rbsR, relA1, rpslL150, lysA1) (Taschner et al., 1988), was grown to steady state in glucose minimal medium containing 6.33 g of K2HPO4.3H2O, 2.95 g of KH2PO4, 1.05 g of (NH4)2SO4, 0.10 g of MgSO4. 7H2O, 0.28 mg of FeSO4.7H2O, 7.1 mg of Ca(NO3)2.4H2O, 4 mg of thiamine, 4 g of glucose and 50 µg of lysine per litre, pH 7.0, at 28°C. Absence of DNA replication was achieved by the addition of rifampicin (300 mg l−1 final concentration) for 90 min at an optical density at 450 nm (OD450) of 0.05 measured with a 300-T-1 spectrophotometer (Gilford Instrument Laboratories) before harvesting of the cells. Filamentation was achieved by a 1:1000 dilution with 1 mg l−1 aztreonam freshly resolved in a saturated Na2CO3 solution at an OD450 of 0.025 and further growth for two mass doublings before harvesting the cells.

Escherichia coli GC7527 ΔmukB::Kan (identical to GC7528, described by Niki et al., 1991) and, when required, E. coli MC4100 were grown at 20°C in rich medium containing 50 g of bactotryptone, 25 g of yeast extract and 15 mmol NaOH per litre (TY), and the osmolarity was corrected to 150 mOsM with NaCl (micro-osmometer; Advanced Instruments). When required, 15 µg ml−1 kanamycin was added to the medium. Cell numbers were monitored using an electronic particle counter (orifice 30 µm). Cultures were considered to be in steady-state growth if the ratio between OD and the number of cells remained constant over time (Fishov et al., 1995).

Overexpression and purification of MukB

Full-length E. coli MukB protein and N- and C-terminal MukB domains [MukBNH2: residues 1–227 (Van den Ent et al., 1999) and MukBCOOH: residues 1197–1486] were expressed as C-terminal His6-tag fusion proteins. Genomic DNA was polymerase chain reaction (PCR) amplified using Pfu DNA polymerase and the following primers: MukBWT (3′-TGACTACCATATGATTGAACGCGGTAAATTTCGCTCA-5′ and 3′-AGTCTACGGATCCACTCGCCTGAGAAGGCGCTTCGTC-5′); MukBNH2 (3′-TGACTACCATATGATTGAACGCGGTAAATTTCGCTCA-5′ and 3′-TGACTACCATATGATTGAACGCGGTAAATTTCGCTCA-5′); MukBCOOH 3′-TGACTACCATATGATTCGTCAGGATATTATTCGT-5′ and 3′-AGTCTACGGATCCACTCGCCTGAGAAGGCGCTTCGTC-5′). PCR fragments were digested with NdeI–BamHI and ligated into cleaved pHis17 vector (B. Miroux, personal communication). C41(DE3) cells (Miroux and Walker, 1996) were transformed and produced MukBWT, MukBNH2 and MukBCOOH after induction with 1 mM IPTG. MukBCOOH is expressed at very high levels as an insoluble protein. Inclusion bodies washed with 50 mM Tris-HCl, pH 7.5, 1% Triton X-100 contained more than 90% MukBCOOH. For large-scale expression of MukBWT or MukBNH2, 12 l of 2× TY containing 100 µg ml−1 ampicillin was inoculated with a 1:100 diluted overnight culture and grown to an OD600 of 0.2–0.3 at 37°C before induction with IPTG. Cells were harvested and frozen in liquid nitrogen. The cells were opened by sonication after the addition of lysozyme in 50 mM Tris-HCl, pH 8.0. After centrifugation, the lysate was loaded onto a 12 ml Ni2+-NTA column (Qiagen). After extensive washing with 300 mM NaCl, 20 mM imidazole, then 50 mM and 75 mM imidazole, all in 50 mM Tris-HCl, pH 6.0, the protein was eluted with 300 mM imidazole, 50 mM Tris-HCl, pH 6.0. Peak fractions were pooled and concentrated before loading onto a Superose 6 HR 10/30 or Sephacryl S200 16/60 size exclusion column (Amersham Pharmacia), equilibrated in 20 mM Tris-HCl, 200 mM NaCl, 1 mM EDTA, 1 mM sodium azide, pH 7.5. The proteins eluted as single peaks. Typically, 4 mg of MukBWT and 30 mg of MukBNH2 protein were obtained.

Antibody production, specificity, purification and Alexa labelling

Antisera against wild-type (MukBWT), against the N-terminal domain (MukBNH2) and against the C-terminal domain of E. coli MukB (MukBCOOH) were generated using 0.1 mg ml−1 1:1 with montenide-diluted purified His6-tagged MukB, His6-tagged MukB(1–225 amino acid residues) and from SDS–PAGE gel excised MukB(1197–1486 amino acid residues), which were injected into rabbits. IgG was purified from rabbit sera by affinity chromatography using a Protein A column (Pharmacia). After application of the antibodies, the column was washed with 20 mM sodium phosphate, pH 7.0, and eluted with 0.1 M glycine-HCl, pH 2.7, at a flow rate of 1 ml min−1.

The antibodies directed against the intact protein detected MukB in the cytosolic extract from exponentially growing E. coli MC4100 and, as expected, not in a similar extract of E. coli GC7527 ΔmukB::Kan, a MukB null mutant (not shown). The antibodies did not recognize proteins in an E. coli membrane fraction (not shown). It should be noted that an average cell contains ≈ 150 MukB molecules (Kido et al., 1996), indicating high affinity of the MukB-binding IgG. An additional protein of about 60 kDa was recognized by the antiserum (not shown). Therefore, the affinity antiserum was purified using the isolated intact MukB protein. The purified IgG was subsequently used to label wild-type cells and mukB null mutant cells to study the subcellular localization of the protein by in situ imunofluorescence microscopy.

Monoclonal mouse IgG1 F168-12 against E. coli FtsZ (Voskuil et al., 1994) was purified similarly using a Protein G column (Pharmacia). Fab fragmentation of antibodies using an Immunopure Fab preparation kit (Pierce) and Alexa labelling of the Fab fragments using Alexa 546 or Alexa 488 protein labelling kit (Molecular Probes) were performed as recommended by the manufacturers. Absorbance spectrum of Alexa-labelled Fab fragments was measured using a Uvikon 923 spectrophotometer (Bio-Tek Instruments). For the MukB affinity purification of the antibodies, 30 µg of MukB protein was run on SDS–PAGE and transferred to a nitrocellulose membrane. The band at proper molecular weight was cut out and incubated overnight at 37°C with an excess of serum antibodies. The antibodies were eluted twice with 0.1 M glycine-HCl, pH 2.7, for 10 min before neutralization of the pH with 1 M Tris-HCl buffer, pH 9.0.

Western blot

Cells were collected at an absorbance at 450 nm of 0.2 and resuspended in PBS containing 0.4 mM Pefabloc, 0.5 mM EDTA, 5 mM dithiothreitol (DTT), 0.02 mg ml−1 DNase and sonicated using a B-12 sonifier (Branson Sonic Power). Unlysed cells were disposed by centrifugation for 5 min at 11 000 g. The samples were centrifuged for 1 h at 284 000 g, supernatants were collected, and protein concentrations were estimated using their 280 nm absorption. Proteins were precipitated using 10% trichloroacetic acid (TCA) final concentration for 15 min, separated from the solvent by centrifugation for 15 min at 15 000 g and washed once in ice-cold acetone. Pellets were dissolved in SDS–PAGE sample buffer and boiled for 5 min. Electrophoresis and immunoblotting were performed as described previously (Laemmli, 1970; Towbin et al., 1979). The membrane was probed with MukB-specific polyclonal IgG. Washing steps were performed according to the Pierce Supersignal Western blotting protocol. The membrane was incubated with horseradish peroxidase-conjugated goat anti-rabbit antibodies (Sanofi Diagnostics Pasteur), and developed with chemiluminescence reagents (Pierce).

Fixation and permeabilization

Cells were fixed in 2.8% formaldehyde (FA) and 0.04% glutaraldehyde (GA) in growth medium for 15 min at room temperature. The cells were collected at 8000 g for 5 min, washed once in PBS (140 mM NaCl, 27 mM KCl, 10 mM Na2HPO4.2H2O, 2 mM KH2PO4, pH 7.2) and subsequently incubated in 0.1% Triton X-100 in PBS for 45 min at room temperature. All subsequent centrifugations were performed at 4500 g. The cells were washed three times in PBS and incubated in PBS containing 100 µg ml−1 lysozyme and 5 mM EDTA for 45 min at room temperature. Finally, the cells were washed three times in PBS.

Immunolocalization experiments

Non-specific binding sites were blocked by incubating the cells in 0.5% (w/v) blocking reagents (Boehringer) in PBS for 30 min at 37°C. Incubation with primary antibodies or Fab fragments conjugated with Alexa 546 or Alexa 488 diluted in blocking buffer was carried out for 60 min at 37°C. The cells were washed three times with PBS containing 0.05% (v/v) Tween-20. When IgG was used, incubation with secondary antibodies, goat anti-rabbit conjugated with Alexa 546 (Molecular Probes), diluted in blocking buffer was carried out for 30 min at 37°C. The cells were washed again three times in PBS, 0.05% Tween-20. The nucleoids were stained with DAPI (4,6-diamino-2-phenylindole) at a final concentration of 0.5 µg ml−1 in PBS. The cells were washed once in PBS and resuspended in PBS.

Microscopy and image analysis

Cells were immobilized on 1% agarose in water slab-coated object glasses as described by van Helvoort and Woldringh (1994) and photographed with a cooled Princeton CCD camera mounted on an Olympus BX-60 fluorescence microscope through a UPLANFl 100×/1.3 oil objective. Images were taken using the public domain program object-image 2.08 by N. Vischer (University of Amsterdam; http://simon.bio.uva.nl/object-image.html), which is based on NIH image by W. Rasband. In all experiments, the cells were first photographed in the phase-contrast mode, then with a DAPI fluorescence filter (U-MWU, excitation at 330–385 nm) and, finally, with the Alexa 546 filter (U-MNG, excitation 530–550 nm) and, when required, also with the Alexa 488 filter (U-MNB, excitation 470–490 nm). The three or four photographs were stacked, and the length of each cell was measured in the phase-contrast image; the DNA segregation was determined in the DAPI image, and the presence of oblong shapes or foci in the fluorescence image. Interactive measurements were performed as ‘structured point collection’ on a Macintosh G3 computer using the public domain program object-image 2.08.

Three-dimensional images were made using a confocal scanning laser microscope (Zeiss) with a Planneofluor 100×/1.3 oil objective and a pinhole of 1.62 µm. The Alexa 546 was excited using a 543 nm line of an argon laser in combination with a LP560 nm filter. Twenty-five 0.1 µm Z-frame scans were performed. Each scan was the result of four times frame averaging to produce a high signal-to-noise ratio. A series of images was made from 0.1 µm fluorescent spheres using the same instrument settings. These images were used to calculate the point spread function (PSF) of the optics with the huygens program (Huygens systems 2.1, Scientific Volume Imaging). The PSF was subsequently used to correct the MukB fluorescence images.


We thank Drs E. M. M. Manders and, P. G. Verschuren for help with the confocal scanning laser microscopy and the huygens program respectively.