MreB is a bacterial actin that plays important roles in determination of cell shape and chromosome partitioning in Escherichia coli and Caulobacter crescentus. In this study, the mreB from the filamentous cyanobacterium Anabaena sp. PCC 7120 was inactivated. Although the mreB null mutant showed a drastic change in cell shape, its growth rate, cell division and the filament length were unaltered. Thus, MreB in Anabaena maintains cell shape but is not required for chromosome partitioning. The wild type and the mutant had eight and 10 copies of chromosomes per cell respectively. We demonstrated that DNA content in two daughter cells after cell division in both strains was not always identical. The ratios of DNA content in two daughter cells had a Gaussian distribution with a standard deviation much larger than a value expected if the DNA content in two daughter cells were identical, suggesting that chromosome partitioning is a random process. The multiple copies of chromosomes in cyanobacteria are likely required for chromosome random partitioning in cell division.
Although many bacteria have a single circular chromosome, linear chromosomes and multireplicon chromosomes have been found in some bacteria (Casjens, 1998). While many of the bacteria with single circular chromosome such as E. coli and C. crescentus have one or two copies of chromosome per cell, some other bacteria with few circular chromosomes have multiple copies of chromosomes during their cell cycles. For example, Azotobacter vinelandii gains multiple copies of chromosomes when grown in rich medium (Maldonado et al., 1994). The presence of multiple copies of chromosomes per cell has also been reported from Borrelia hermsii (Kitten and Barbour, 1992) and Mirococcus radiodurans (Hansen, 1978). The cyanobacteria that have been studied so far have multicopy chromosomes at all stages of their cell cycles (Herdman et al., 1978; Labarre et al., 1989; Tandeau de Marsac, 1994). Although these bacteria must also faithfully replicate their DNA and separate their chromosomes into daughter cells in cell division, it is not known if MreB plays the same roles in the chromosome segregation of these organisms.
In this study, we chose a filamentous cyanobacterium, Anabaena sp. PCC 7120, to study the roles of MreB in regulating cell morphology and chromosome segregation. The filaments of Anabaena sp. PCC 7120 are composed of ellipse-shaped cells. Under nitrogen deficiency condition, some vegetative cells differentiate and become heterocysts (Haselkorn, 1978; Wolk et al., 1994; Zhang et al., 2006). Like other cyanobacteria, Anabaena sp. PCC 7120 performs oxygenic photosynthesis. Its photosynthetic electron transfer chain is located on thylakoid membranes, which are often arranged parallel to cytoplasmic membranes (Gantt, 1994). Its DNA is localized near the centre of cytoplasm, inside the thylakoid membranes. Here we report that inactivation of mreB genes led to a dramatic change in cell shape in Anabaena sp. PCC 7120. However, the strain lacking mreB remained filamentous and had a same growth rate as the wild type. Chromosome segregation was normal in the mutant. Our results suggest that the cyanobacteria, which are known to have multicopy chromosomes, may adapt an MreB-independent mechanism of chromosome segregation.
Inactivation of mreB of Anabaena sp. PCC 7120
Unlike B. subtilis, which has three mreB homologues, Anabaena sp. PCC 7120 has one mreB and no other mreB-homologue has been found through similarity search of its genome. Figure 1A shows a schematic drawing of mreBCD and the inactivation of the mreB gene from Anabaena sp. PCC 7120. The mreB gene was interrupted with a cartridge containing a gene encoding erythromycin resistance. Figure 1C presents the result of a Southern blot showing that the mutant strain had a correct insertion of the cartridge and no wild-type mreB gene was present. A 6.4 kb fragment in the wild-type total genomic DNA digested with HincII was identified with the mreB probe; this band was absent from total DNA of the mutant digested with HincII. The mutant instead had a 2.4 kb fragment that hybridized with the probe. The hybridization pattern was as expected from the construction and it was concluded that the mreB gene was inactivated in the mutant strain, which was named Mbd320. Various mre gene constructions that were used to complement Mbd320 were shown in Fig. 1B. The transcription of three mre genes was analysed with Northern blot (Fig. 1D) and quantitative polymerase chain reaction (PCR) (Fig. 1E). Two transcripts were detected in the wild type when hybridized with a probe containing part of the mreC gene: one was 2.7 k and the other 1.6 k. The large transcript was absent in Mbd320 but the small one was present in the mutant. When a probe for mreB was used in Northern blot, we also observed two transcripts, a 2.7 k band, which was same in length as that detected with the mreC probe, and a 1.3 k band, which was not large enough to contain full-length mreC. Neither band was detected in Mbd320. These results suggest that mreBCD genes are transcribed as an operon in Anabaena sp. PCC 7120 (Fig. 1A) and that an additional promoter for mreCD is present. The non-coding sequence between mreB and mreC has 98 bp and there is a potential promoter in this region based on the consensus −10 box sequence (TAGAAT, the consensus residues are underlined) of cyanobacterial promoters (Su et al., 2006). Quantitative PCR (Fig. 1E) confirmed that the mreB transcript was absent in Mbd320 while the transcript of mreCD was present. The amount of mreCD transcript was only a third of that observed in the wild type, indicating that the larger transcript contributed more to the mreCD transcript abundance.
Figure 2A and B show that the wild-type cells of Anabaena sp. PCC 7120 had ellipse shape and connected to one another in such a way that the cell's long axes were the same as that of the filaments. The cells of Mbd320, which were also ellipse-shaped, were connected to one another such that the cell's axes were perpendicular to the axes of the filaments (Fig. 2C and D). The cells of Mbd320 were nearly twice as large as the wild-type cells (Table 1). This is likely due to an increased size of cell septa of Mbd320. The cell shape of Mbd320 remained the same without the presence of the antibiotic in growth medium, confirming that the mutant had no wild-type mreB gene left. Complementation of Mbd320 with various mre genes under control of the petE promoter (Fig. 1B), which is inducible by copper (Buikema and Haselkorn, 2001), was performed. When Mbd320 was complemented with the entire mreBCD operon on a shuttle plasmid (strain Mbd321), its cell morphology was very similar to that of the wild type when grown in the presence of 0.2 μM copper (Fig. 2E). No complementation was found with constructions that expressed mreB, mreBC and mreCD in the presence of 0.2 μM and 2 μM copper (data not shown). When grown in the presence of 2 μM copper, which does not cause any morphological change of Anabaena sp. PCC 7120 (Buikema and Haselkorn, 2001), Mbd321 formed long, irregular cells (Fig. 2F). One phenotype that we observed in Mbd320 besides its cell shape was that the cells were more sensitive to lysozyme (Fig. S1), indicating that there was some defect of the Mbd320 cell walls. However, under a normal condition, the growth of Mbd320 and the wild type was similar. Figure 3 shows the growth curves of the wild type and Mbd320 in the presence of combined nitrogen. The doubling times for the wild type and Mbd320 based on optical density were 21 h and 21.3 h respectively (Fig. 3A). The doubling times for the wild type and Mbd320 based on cell numbers were 20.6 h for the wild type and 21.1 h for the mutant (Fig. 3B), demonstrating that Mbd320 has a similar growth rate to the wild type.
Table 1. Comparison of the cells from the wild type and Mbd320.
The dimension parallel to the filaments of Anabaena sp. PCC 7120.
The dimension perpendicular to the filaments.
The relative DNA content as determined DAPI-staining of a given cell (cell n) over that of its neighbour cell (cell n + 1).
Chromosome copy numbers in Mbd320 and the wild type
Recent evidence shows that mreB is required for chromosome segregation in E. coli, C. crescentus and B. subtilis. To investigate the role of MreB in chromosome segregation of Anabaena sp. PCC 7120, the copy number of chromosome per cell in both strains was determined by quantitative PCR. Three separate loci on Anabaena sp. PCC 7120 genome were amplified and their copy numbers were compared with a plasmid with known concentration. The copy number of chromosome per cell (cell number was determined by cell counting) is the average of the three measurements (Table 1). Like Synechocystis sp. PCC 6803 (Labarre et al., 1989), Anabaena sp. PCC 7120 has many chromosomes per cell. On average, the cells of the wild type and Mbd320 contained 8.2 ± 1.6 and 10.2 ± 2.1 copies of chromosomes per cell under our growth conditions. To study if inactivation of mreB resulted in any impairment of chromosome segregation, we used 4,6-diamidino-2-phenylindole (DAPI) to stain the filaments of both strains of Anabaena sp. PCC 7120. The emission with wavelength shorter than 600 nm was measured and the spectra showed the DAPI-dependent fluorescence emission peak at 460 nm (Fig. S2). The fluorescence images shown in Fig. 4A and B contained the fluorescence of both photosynthetic pigments and the DAPI-stains. It is evident that the thylakoid membranes with photosynthetic pigments from both strains were located at the periphery of the cells and the DAPI-stains were located at cell centres. When only fluorescence with wavelength shorter than 480 nm was measured, the images of blue fluorescence from the wild type (Fig. 4C) and Mbd320 (Fig. 4D) represented the DAPI-stains of chromosomes. It is evident that the DAPI-stains showed a relatively even distribution along the filaments in both strains. When the relative DNA content determined by DAPI-stain from a cell (cell n) on filaments was plotted against that of next cell (cell n + 1) (Fig. 4H), both the wild type and Mbd320 had a similar pattern of distribution. The ratios obtained from both strains (Fig. 4C and D), which reflected the relative fluctuation of cellular DNA content along the filaments, distributed along a line with a slope of 1. Although scattering is observed in both distributions, the average ratios were very close to 1 and their standard deviations were nearly identical (Table 1), suggesting that the lack of mreBCD did not lead to irregular distribution of cellular DNA content along the filaments.
The DAPI-stain mentioned above relies on the addition of the dye to cells externally, which could introduce staining heterogeneity along the filaments. We adapted the GFP–ΔParB/parS system of P1 plasmid (Li and Austin, 2002) to observe the chromosomes in Anabaena sp. PCC 7120 and it was possible because there is no sequence in the genome of Anabaena sp. PCC 7120 that shows significant homology to the P1 parS sequence. The GFP–ΔParB/parS system depends on aggregate formation of GFP–ΔParB fusion protein on the parS sequence. The P1 parB–parS sequence was integrated into a docking site of the Anabaena sp. PCC 7120 genome as gfp–ΔparB–parS fusion under control of the petE promoter by double recombination. Southern hybridization confirmed that every chromosome contained one copy of the gfp–ΔparB–parS fusion (data not shown). As shown in Fig. 4E and F, GFP–ΔParB can be observed in the filaments of the wild type and Mbd320 and they formed aggregates due to the presence of multiple chromosomes in each cell and individual spots of GFP were often not resolved. Occasionally, however, some individual GFP–ΔParB/parS aggregates could be observed in a few cells with a confocal microscope and the one shown in Fig. 4G indicates there were approximately 10 chromosomes in that cell (indicated by an arrow). When the relative GFP fluorescence from a cell (cell n) on filaments was plotted against that of next cell (cell n + 1) (Fig. 4I), both the wild type and Mbd320 had a similar pattern of distribution to that obtained by DAPI-stains (Fig. 4H). The ratios obtained from both strains also distributed along a line with a slope of 1. The scattering of the distribution is less in the GFP–ΔParB/parS plot than in the DAPI-stain plot.
Several factors such as positions in cell cycle and chromosome copy numbers could contribute to the fluctuations of DNA content shown in Fig. 4H and I. Cells in a filament were not synchronized in cell cycle and those cells prior to cell division were likely to have more DNA than the cells that just finished cell division. To make a more accurate measurement, we chose pairs of daughter cells that just finished cytokinesis to measure their relative DNA content. Figure 5A and B show the distribution of the DAPI-stained DNA of the cell pairs from the wild type and Mb320, respectively, and both histograms show a pattern of Gaussian distribution. When ratios of the DNA content of two daughter cells were plotted (Fig. 5C), the histograms also showed a Gaussian distribution. The GFP fluorescence from GFP–ΔParB/parS also reflects the DNA content because it is dependent upon the copy number of parS on chromosomes. The GFP fluorescence intensity of the cell pairs that just finished cell division was determined and ratios were plotted (Fig. 5D). The distributions of the fluorescence ratios in both the wild type and Mbd320 were Gaussian with standard deviation of 0.286 and 0.278 respectively. A strain of Anabaena sp. PCC 7120 expressing gfp under a tac promoter of E. coli was used as a control. The GFP fluorescence from pairs of cells that just finished cell division was determined and the ratios were plotted (Fig. 4D). Although it is also in Gaussian distribution, the peak is much narrower than those obtained with GFP–ΔParB/parS system. The standard deviation of the distribution is 0.201, which is significantly different from the values obtained in the wild type and Mbd320 (N > 0.99). These results suggest that the DNA content in two daughter cells were not always identical and that some random event might be involved in the partitioning of chromosomes into the two daughter cells in cell division of Anabaena sp. PCC 7120.
Effect of mreB inactivation on cell division and thylakoid organization
Although Mbd320 showed a big change in cell morphology, its cell division was not affected by the lack of mreB. We used a ftsZ–gfp fusion to localize the cell division Z-ring in both the wild type and Mbd320. FtsZ is a bacterial tubulin that forms a ring at the place where a septum is to be formed (Rothfield et al., 1999). The fluorescence images of FtsZ–GFP in both strains are shown in Fig. 6. It is clear that the FtsZ-rings in both the wild-type cells (Fig. 6A) and Mbd320 (Fig. 6B) were localized near the centre of a dividing cell as reported by Sakr et al. (2006), suggesting that the lack of mreB did not result in an impairment of septum placement. Figure 6C and D show images of Mbd320 and the wild type obtained from thin-sectioned electron microscopy respectively. The cell walls between two cells in Mbd320 were complete. The thylakoid membranes were distributed parallel to cytoplasmic membranes in both strains, indicating that mreBCD was not specifically involved in thylakoid organization.
Localization of MreB in Anabaena sp. PCC 7120
It has been demonstrated that MreB forms spiral fibres underneath the cytoplasmic membranes and this organization of MreB was suggested to be critical in control of cell shape. The localization of MreB in Anabaena sp. PCC 7120 was studied by complementing Mbd320 with a gfp–mreB–mreCD fusion gene operon. The shape of the complemented cells was similar to that of the wild type (Fig. 7B), demonstrating that the gfp–mreB fusion was functional. The green fluorescence from the gfp–mreB fusion product was mostly located at poles of both the vegetative cells and heterocyst (Fig. 7). In some cells, the green fluorescence could be observed underneath the cell membranes as fibre bundles, indicating MreB assembly could be a dynamic process and there could be more than one form of the MreB complex.
Anabaena sp. PCC 7120 is capable of forming heterocysts under nitrogen deprived conditions. Because heterocysts of Anabaena sp. PCC 7120 have different cell morphology as compared with vegetative cells, we investigated how mreB gene was expressed in heterocysts. Figure 8A and B show that both the wild type and Mbd320 formed long filaments with normal heterocyst patterns. Mbd320 was also capable of growing normally in the absence of combined nitrogen (data not shown), suggesting that its heterocysts were able to fix nitrogen. A gfp gene under control of the mreB promoter on a shuttle vector was transformed into the wild-type Anabaena sp. PCC 7120 and the fluorescence image from GFP (Fig. 8C) shows that the mreB promoter was upregulated because the GFP fluorescence from differentiated cells were brighter than that of the vegetative cells. The red fluorescence image of photosynthetic pigments is shown in Fig. 8D.
MreB is a bacterial actin that plays a critical role in chromosome segregation in C. crescentus (Gitai et al., 2005) and E. coli (Kruse et al., 2006). The separation of oriC is dependent on MreB in both organisms. It has been suggested that either MreB could provide a track for motor-like proteins to move chromosome or its polymerization could provide the force for DNA movement (Moller-Jensen et al., 2003; Graumann and Defeu Soufo, 2004; Kruse et al., 2006). Little is known about the mechanism of chromosome segregation in bacteria that contain multiple copies of chromosome such as cyanobacteria. The roles of mreBCD in chromosome segregation in Anabaena sp. PCC 7120 were investigated in this study. The genes of mreB, mreC and mreD of Anabaena sp. PCC 7120 are organized like an operon as in E. coli (Wachi et al., 1989), B. subtilis (Levin et al., 1992) and some other rod-shaped cyanobacteria (http://www.kasusa.or.jp/cyano). Northern blot shows that these three genes are transcribed as an operon. Northern blot and quantitative PCR also demonstrate that mreCD could be transcribed from a promoter located downstream of the mreB gene, similar to the transcription of the mreBCD genes in E. coli (Wachi et al., 2006). Although inactivation of mreB in the photosynthetic bacterium Rhodobacter sphaeroides was not possible (Slovak et al., 2005), the mreB gene in Anabaena sp. PCC 7120 was successfully inactivated as confirmed by Southern hybridization (Fig. 1) and by the drastic change in cell shape in the mutant (Fig. 2).
It was somewhat surprising that the lack of mreB did not result in any abnormal distribution of DNA in two daughter cells of cell division in Anabaena sp. PCC 7120. If chromosome segregation was impaired in Anabaena sp. PCC 7120 by inactivation of mreB, it would be expected that some cell death would occur and fragmentation of filaments should be observed. However, Mbd320 had long filaments as the wild type, and the growth rate of Mbd320 was same as that of the wild type, suggesting that no cell death occurred as a result of mreB inactivation. In B. subtilis deletion of mreB did not result in an apparent defect in chromosome segregation if it was grown in the presence of magnesium and sucrose (Formstone and Errington, 2005). It is likely that an mreB homologue could provide similar roles of mreB in chromosome segregation as suggested by Kruse et al. (2006). Because Anabaena sp. PCC 7120 has no other mreB homologue in its genome, we conclude that chromosome segregation of Anabaena sp. PCC 7120 is not dependent upon MreB. The same conclusion is obtained in Stryptomyces coelicolor that shows no requirement of MreB for its vegetative growth (Mazza et al., 2006).
One unique feature of cyanobacterial cells is that they have thylakoid membranes, which are the sites of photosynthetic electron transfer, ATP synthesis and phycobilisome attachment. The thylakoid membranes are often arranged in a concentric fashion and connections between thylakoid membranes and cytoplasmic membranes are also present (Gantt, 1994). The DNA molecules are located at the centre of the cyanobacterial cells (Fig. 4). The organization of thylakoid membranes could be a big hindrance if a chromosome is required to move along the track of MreB fibre as proposed for C. crescentus (Gitai et al., 2005, Margolin, 2005).
It has long been recognized by researchers that there are many copies of chromosomes per cell in the cyanobacteria. The reason for the presence of multiple copies of chromosomes in the cyanobacteria is little studied and the mechanism for DNA replication in cyanobacteria is poorly understood. Although components of DnaA-dependent machinery for initiation of DNA replication are present, the dnaA gene appears not to be essential for the cyanobacterium Synechocystis sp. PCC 6803 (Richter et al., 1998). Another striking feature of the cyanobacterial genomes is that no apparent oriC sites could be found based on DNA asymmetry (Kaneko et al., 2001). The mechanism of DNA replication initiation in cyanobacteria is therefore unknown. Although it has been shown that cell division is coupled to DNA replication (Herdman and Carr, 1971; Yoshida et al., 2005) except for hormogonium differentiation in which cell division takes place without DNA replication (Tandeau de Marsac, 1994), little is known at present how the cyanobacterial cells separate many copies of chromosomes into daughter cells in cell division. We used two systems to measure the relative DNA content in two daughter cells after cell division: DAPI-stain and GFP–ΔParB/parS system. The latter system has been successfully used to locate individual chromosome in several bacteria such as E. coli (Li and Austin, 2002). In our system, the GFP fluorescence intensity from GFP–ΔParB between two daughter cells would depend on the copy number of parS in each cell. The evidence obtained by both methods shows that chromosomes of Anabaena sp. PCC 7120 are not precisely separated into daughter cells because the DNA content in the population of daughter cells after cell division in both the wild type and Mbd320 was not always identical. The ratios of DNA content of the two daughter cells after cells division had a Gaussian distribution with standard deviations significantly larger than the control (Fig. 5). These results suggest that chromosomes in Anabaena sp. PCC 7120 are partitioned into two daughter cells randomly. In bacteria, plasmids without a partitioning function (Par–) are partitioned randomly and they are often present in many copies per cell (Nordstrom and Gerdes, 2003). Random partition of chromosomes would require multiple copies of chromosomes per cell so that no cell death would occur because of random partitioning, implying that the presence of multicopy chromosome in cyanobacteria is required for maintaining chromosome after cytokinesis by random partitioning.
Cell division was unaffected by the lack of mreBCD based on the observation of the formation of FtsZ-ring and cell septa in Mbd320 (Fig. 6). Therefore, these results raise the possibility that the Mre proteins do not participate in the Z-ring and septum formation. Supporting this suggestion is that the spherical cyanobacterium Synechocystis sp. PCC 6803 apparently lacks mreB. In bacteria, the Z-ring placement is determined by MinCDE system and nucleoloid occulation (Rothfield et al., 2005). Recently, a MipZ-dependent determination of Z-ring placement has been reported in C. crescentus (Thanbichler and Shapiro, 2006). Because of the presence of thylakoid membranes and the central location of chromosomes in Anabaena sp. PCC 7120, it is unlikely that the Z-ring placement was determined by nucleoloid occulation mechanism in this organism. We searched the cyanobacteria genomes and found no homologue of MipZ of C. crescentus while MinC, MinD and MinE are present in the cyanobacteria. Therefore, the site of septum formation in cell division in cyanobacteria is probably decided by the MinCDE system. The heterocysts and their pattern are normal in Mbd320 under a nitrogen-limiting condition, suggesting the cell–cell communication along the filaments remains same as that of the wild type. It is interesting to note that the mreB promoter was upregulated in differentiated cells, indicating that MreBCD proteins could play some role in heterocysts in Anabaena sp. PCC 7120.
The roles of mreBCD in determination of cell shape are similar to that reported in some other bacteria (van den Ent et al., 2001b; Jones et al., 2001). The GFP–MreB, which is functional in Anabaena sp. PCC 7120, is mostly located at poles of the cells and it suggests that MreB could inhibit septum synthesis so that rod-shaped cells are formed as in other rod-shaped bacteria (Stewart, 2005). Another possible role of MreB in Anabaena sp. PCC 7120 could be the determination of cell polarity as in C. crescentus (Gitai et al., 2004) based on its location (Fig. 7) although further research is needed to know whether the location of MreB at cell poles determines the cell polarity or it is a consequence of cell polarity determined by other factors. The bundle-like form of GFP–MreB in some cells of Anabaena sp. PCC 7120 suggests that MreB supramolecular structure could be as dynamic as reported in other bacteria (Defeu Soufo and Graumann, 2004; Gitai et al., 2004). Complement studies showed that mreB gene alone could not complement Mbd320 while a mreBCD operon could. Together with the fact that the amount of mreCD transcript was much reduced in Mbd320, these results indicate that the amount of the mreCD transcripts could also be critical to the determination of cell shape. Overexpression of mreBCD (Fig. 2) caused long and irregular cells while the cell shape of Mbd320 was not changed when complemented by mreCD, indicating that the effect of overexpression of mreCD required the presence of MreB. The results presented here suggest that all three mre genes were involved in cell shape determination in Anabaena sp. PCC 7120 as reported from other bacteria such as B. subtilis (Leaver and Errington, 2005).
Strains and culture conditions
Both the wild type and mutant strains of Anabaena sp. PCC 7120 were grown in BG11 medium (Rippka et al., 1979) with or without nitrate at 28°C under cool white fluorescent light at 70 μmol m−2 s−1. The cultures were bubbled gently with air plus 1% CO2. The growth was measured by monitoring optical density at 730 nm or by counting the cell numbers. To induce expression of genes under control of the petE promoter of Anabaena sp. PCC 7120, copper-free media were prepared according to Buikema and Haselkorn (2001) and induction of the genes was performed by addition of appropriate concentration of copper. Heterocyst differentiation was induced by transfer the cells to a medium without combined nitrogen. E. coli strains were grown in LB medium supplied with appropriate antibiotics. E. coli strain DH5α was used for all routine cloning purpose and strain BL21(DE3) was used for overproduction of recombinant proteins.
Gene inactivation and complementation
All enzymes were purchased from Promega (Beijing) and used according to the instructions. To construct an mreB mutant (Mbd320), a DNA fragment containing mreBCD operon was amplified by PCRs with primers P1 and P2 (these two and all other primers used in PCR are listed in Table S1), using the total genomic DNA as template. The PCR-generated fragment was cloned in pGEM-T vector (Promega) to generate plasmid pTv-BCD. The plasmid was inversely amplified by PCR with primers P3 and P4 and the generated fragment was ligated with a blunt-ended fragment containing eryR cartridge. The resulted plasmid pBCD-em was digested with EcoRI and the generated fragment was then cloned into pRL277 (Elhai and Wolk, 1988) for transformation of Anabaena sp. PCC 7120 through conjugation. Segregation of the mutant mreB was confirmed by Southern hybridization performed as described by Huang et al. (2004). Total genomic DNA of Anabaena sp. PCC 7120 was isolated with the E.Z.N.A. Planet DNA miniprep kit (Omega Bio-tek, Beijing) and it was digested with HincII. The generated fragments were separated by a 1.0% agarose gel and transferred onto a nitrocellulose paper before hybridized with a DNA probe, which was then synthesized by random primer extension. The template used for probe synthesis was the coding region of the mreC from Anabaena sp. PCC 7120 amplified with primers P5 and P6. Total RNA isolation and Northern hybridization were performed according to Huang et al. (2004). The probe of mreC used in the Northern hybridization was amplified by PCR using primers P5 and P6. Quantitative PCR for analyses of the expression of mreB, mreC and mreD were performed as described by Huang et al. (2004). The primers used for quantitative PCR of mreB, mreC and mreD were Q7/Q8, Q9/Q10 and Q11/Q12 respectively. For complementary study of the Mbd320, the mreBCD, mreBC, mreB and mreCD were amplified with PCR with primers P7/P9, P7/P6, P7/P8 and P5/P9 respectively. The amplified genes were cloned into pAM505 (Yoon and Golden, 1998) under control of the petE promoter as described by Zhao et al. (2005). The genes were induced with either 0.2 μM CuSO4 for normal expression or 2 μM CuSO4 for overexpression.
Cellular localization of MreB
For determination of the spatial pattern of mreB expression, the mreB promoter was amplified with primers P10 and mreP11 and the amplified fragment was digested with SalI and NcoI before it was cloned into the pET30a, generating pPmreB. A gfp gene was amplified with PCR using primers P12 and P13 using the plasmid pAM1951 (Yoon and Golden, 1998) as template. The fragment was digested with BspHI and cloned into pPmreB under control of the mreB promoter. The entire fragment was then cloned into the pAM505 and transformed into Anabaena sp. PCC 7120 as described (Zhao et al., 2005). To determine the cellular location and supramolecular structure of the MreB, the mreBCD operon was amplified with primers P7 and P9 and cloned into pAM505 as described above. The amplified gfp gene described above was then fused with mreB on pAM505 in frame. The resultant plasmid pAM505-gfpBCD (Fig. 1) was transformed into Mbd320. To construct an ftsZ–gfp fusion for FtsZ-ring observation, the petE promoter was first amplified with P15 and P16 and the fragment was digested with BglII and XbaI. It was then ligated to pET-3d that was digested with the same enzymes. The petE promoter was then digested with BglII and NcoI and ligated into pET-30a, generating pET30a-PpetE. The ftsZ gene from Anabaena sp. PCC 7120 (Deherty and Adams, 1995) was amplified with primers P17 and P18 and cloned into the EcoRI site of pET30a-PpetE, generating plasmid pPpetE-FtsZ. The gfp gene was fused to the ftsZ gene on pPpetE-FtsZ as in construction of mreB–gfp fusion, generating pPpetE-ftsZ–gfp. The plasmid was digested with BglII and SalI and the generated fragment containing ftsZ–gfp under control of the petE promoter was cloned into pAM505 and transformed into the wild type and Mbd320.
The chromosomal copy numbers of Anabaena sp. PCC 7120 strains were determined by quantitative PCR using total DNA as template. Three genes were chosen for the amplification: all0001, hetR (Buikema and Haselkorn, 1991) and alr3807, which are located at the positions of 0 Mb, 2.3 Mb and 4.6 Mb on the 6.4 Mb genome map of Anabaena sp. PCC 7120 (Kaneko et al., 2001) respectively. The results obtained from quantitative PCR were averaged. The primers used for the amplifications of internal fragments of all0001 (Q1 and Q2), hetR (Q3 and Q4) and alr3807 (Q5 and Q6) are given in Table S1. Plasmids containing entire coding sequences of all0001 and alr3807 and the plasmid pHetRΩ (Dong et al., 2000) were as standard in quantitative PCR as described by Huang et al. (2004). The cell number in a culture from which total DNA was isolated for quantitative PCR was determined by cell counting.
The GFP–ΔParB/ParS system (Li and Austin, 2002) was used to observe individual chromosome in Anabaena sp. PCC 7120 as follows. The ΔparB–parS (ΔparB encodes a truncated ParB that had its first 30 amino acid residues deleted) was obtained by amplifying the P1 plasmid using primers P19 and P20 and cloned into p3dPpetE (Zhao et al., 2005). A gfp gene was inserted into the NcoI site so that a gfp–ΔparB fusion gene was obtained as the gfp–mreB fusion was constructed as mentioned above. The fragment containing gfp–ΔparB–parS under control of the petE promoter was amplified by PCR using primers P21 and P22 and inserted into the docking site of plasmid pRL-277d, which was a derivative of pRL277 (Elhai and Wolk, 1988) for insertion of foreign DNA sequences. The plasmid pRL277d contained the 6 kb fragment between alr3952 and all3953 of Anabaena sp. PCC 7120 and it had a kan cartridge inserted in the middle region of the 6 kb sequence. The EcoRI site was used for the insertion of the fragment containing gfp–ΔparB–parS under control of the petE promoter. A gfp gene under control of a tac promoter was also inserted into the EcoRI site of pRL277d. Both plasmids were transformed into Anabaena sp. PCC 7120 through conjugation.
Microscopy and image acquisition
Bright field and fluorescence microscopy was performed with an Olympus BX51 microscope equipped with a CCD camera as described by Huang et al. (2004). DAPI-stain was performed according to Formstone and Errington (2005) with a DAPI concentration of 1 μg ml−1. Observation of fluorescence from photosynthetic pigments and GFP was performed with an excitation light at 466 nm according to Zhao et al. (2005). DAPI-stained DNA was excited with a UV light at 366 nm. The images of DAPI-stains and GFP–ΔParB aggregates on parS sequence were stored digitally and analysed with Image Pro Plus software according to the supplier (Apogee Auburn, CA). For observation of FtsZ-ring, a confocal laser-scanning microscope was used according to the instructions of the supplier (Leica, Germany).
Electron microscopy was performed with a JEM-1010 electron microscope (Joel, Tokyo, Japan). Sample preparation and thin sectioning were carried out according to Ohki and Fujita (1992).
We thank Professor J. Golden for kindly providing plasmids and Professor P. Wolk for his discussion and comments on cyanobacterial chromosome copy numbers. We thank Yuxian Zhu, Yiping Wang and Hongya Gu for discussion and comments on the manuscript and Dr H. Sang for her assistance on confocal microscopy. This research is supported by the National Science Foundation of China (30230040) and The Ministry of Science and Technology of China (01CB108903).