SMC (structural maintenance of chromosomes) proteins are highly conserved and present in eukaryotes, bacteria and archaea. They function in chromosome condensation and segregation and in DNA repair. Using an insertion vector containing the pac gene for resistance to puromycin, we have created an insertion in the smc gene of Methanococcus voltae. We used epifluorescence microscopy to examine the cell and nucleoid morphology, DNA content and metabolic activity. This insertion causes gross defects in chromosome segregation and cell morphology. Approximately 20% of mutant cells contain little or no DNA, and a subset of cells (≈ 2%) IS abnormally large (three to four times their normal diameter) titan cells. We believe that these titan cells indicate cell division arrest at a cell cycle checkpoint. The results confirm that SMC in archaea is an important player in chromosome dynamics (as it is in bacteria and eukaryotes).
Faithful segregation and organization of newly replicated chromosomes is vital in all cells. Although the molecular machinery involved in chromosome dynamics in all the domains of life is only poorly understood, members of the structural maintenance of chromosomes (SMC) family of proteins perform an important role. SMC proteins function in several complexes involved in the dynamic activities of chromosomes during the cell cycle, such as cohesion, condensation, dosage compensation and recombinational DNA repair (for recent reviews, see Graumann, 2001; Hirano, 2002; Jessberger, 2002). SMC complexes have generated extensive interest (Gruber et al., 2003; Hagstrom and Meyer, 2003; Milutinovich and Koshland, 2003) because of their role in almost all aspects of chromosomal biology in both prokaryotes and eukaryotes.
In eukaryotes, members of the SMC protein family are classified into three groups by their ability to form specific SMC heterodimer complexes: SMC1:SMC3 (cohesins), SMC2:SMC4 (condensins) and SMC5:SMC6 (DNA repair) (Hirano, 2002; Jessberger, 2002). These heterodimers also associate with other non-SMC proteins to form fully functional SMC holocomplexes (Hirano, 2002).
Most prokaryotes have a single smc gene (Soppa, 2001). However, a number of organisms with complete genomes (including Rickettsia prowazekii, Helicobacter pylori, Haemophilus influenzae, Chlamydia trachomatis and Methanobacterium thermoautotrophicum) and several major phyla (γ-proteobacteria and crenarchaeotes) contain no smc genes (Soppa, 2001). Among bacteria, SMC proteins have been best studied in Bacillus subtilis and Caulobacter crescentus. In prokaryotes, SMC proteins form homodimers and are partially responsible for proper chromosome condensation and migration into daughter cells. B. subtilis smc deletion mutants show substantial (>10%) anucleate cell production, decondensed chromosomes, temperature-sensitive growth, a reduced growth rate at permissive temperatures in a rich growth medium and disruption of replication origin positioning (Britton et al., 1998; Graumann et al., 1998; Moriya et al., 1998; Britton and Grossman, 1999; Graumann, 2000; Graumann and Losick, 2001). Similarly, smc null mutations in C. crescentus give rise to temperature-sensitive growth and abnormal nucleoid morphology. However, a high percentage of anucleate cells is not observed (<0.1%) in this species (Jensen and Shapiro, 1999; 2003). Escherichia coli, a γ-proteobacterium, lacks any smc genes, but does contain a structurally and functionally similar (yet genetically different) protein for chromosome segregation, called MukB (Niki et al., 1991; Melby et al., 1998). Deletion of mukB leads to phenotypes comparable to that of B. subtilis with an smc disruption (Niki et al., 1991; Melby et al., 1998).
Here, we report the existence of an smc gene in M. voltae and show that it is essential for proper chromosome condensation and segregation. In this study, an smc mutant strain was constructed by replacing the 3′ end of the smc gene with an integration vector containing an antibiotic resistance gene. In the absence of a functional gene product, mutants show a high rate of anucleate cell production and a subset of cells that is extremely enlarged with irregular morphology.
Analysis of the smc gene of M. voltae
The smc gene of M. voltae codes for a protein of 1199 amino acids. This protein is 53% identical to the SMC homologue of Methanocaldococcus jannaschii. It has all the common elements of SMC proteins (described above). There are no significant open reading frames (ORFs) upstream, but there is an ORF coding for a thermonuclease homologue downstream. This ORF is coded for on the opposite strand and would not be disrupted by an insertion in the smc gene. Upstream, there are no ORFs > 100 amino acids until an ORF homologous to aminotransferases, also coded for on the other strand. Thus, it appears that smc is transcribed as a single gene.
Analysis of transformants
After transformation of M. voltae with an insertion vector containing an internal portion of smc, 10 streak-isolated, puromycin-resistant colonies were picked from bottle plates and inoculated into liquid medium. We determined by Southern blot analysis using two different smc-specific probes that one of 10 transformants picked had an insertion in the expected part of the smc gene. The two probes used were SMC-I and SMC-N. Probe SMC-I is complementary to the internal fragment cloned into the EcoRI site of the insertion vector. This fragment hybridizes a 3 kbp band from wild-type M. voltae(Fig. 1). EcoRI digestion of nine puromycin-resistant transformants showed a 3 kbp band plus another band of variable size, indicating that the vector inserted elsewhere in the chromosome (data not shown). These nine transformants were assumed to have the pSLI-1 vector integrated at non-smc sites and were not characterized further. One puromycin-resistant mutant, designated M6, lacked the 3 kb band found in the wild-type sample, but contained three other bands that bound probe SMC-I (Fig. 1). These three EcoRI fragments that hybridized SMC-I are explained by tandem insertions (Fig. 1). The three bands are: (i) the internal smc fragment cloned into the EcoRI site of the pac/pUC vector; (ii) the disrupted 5′ end of the smc gene; and (iii) the disrupted 3′ end of the smc gene. To confirm disruption of the smc gene, a second probe (SMC-N) was used that is complementary to the N-terminal region of the smc gene. This probe is not complementary to any part of the pSLI-1 integration vector. NcoI digestion of wild-type genomic DNA produces a single hybridizing band > 12 kbp. NcoI does not cut inside the smc gene, but there are two NcoI sites in pSLI-1. Using NcoI, SMC-N and M6 DNA in a Southern blot produces a single band of ≈ 5 kbp, confirming that the smc gene is disrupted in clone M6 (Fig. 2). Figure 1A also shows a schematic of the region around smc and a schematic of the region after the insertion. The insertion results in a truncated product that ends ≈ 75 amino acid residues C-terminal to the signature SMC hinge motif G(X)6G(X)3GG. The insertion would disrupt the smc gene such that a truncated 85 kDa protein would be produced instead of the full-length 138 kDa protein.
Nucleoid and cell morphology in smc::pac mutants
We expected an insertional inactivation of smc in M. voltae to have defects in nucleoid compaction and distribution. We observed cells using phase-contrast illumination, DAPI-specific fluorescence illumination and F420 autofluorescence. Methanogenic archaea, as a result of their unique metabolism, will autofluoresce after UV illumination (Doddema and Vogels, 1978). We used this phenomenon to observe the integrity of cells. Bright autofluorescence indicates active proteins and cofactors in the cells before fixation and staining. In particular, intact cells could easily be distinguished from partially or fully lysed cells (ghost cells). Methanococcus cells are easily lysed by handling (Koval and Jarrell, 1987). Thus, we could differentiate cells that had no or little DNA as a result of leakage and cells that had no DNA because of chromosomal segregation defects. The epifluorescence filters used did not allow the DAPI and autofluorescence signals to overlap.
Under phase contrast, wild-type cells appear to be a consistent size (0.8–2.5 µm in diameter) with the expected cell morphology. DAPI staining of wild-type cells reveals that DNA is present in almost all cells, and nucleoid regions can easily be discerned (Fig. 2A), including predivisional segregated nucleoids. The fraction of anucleate wild-type M. voltae was <1% (Table 1). Others have noted in M. jannaschii that anucleate cells are observable, depending on the growth state of the cells and on the fixation method (Malandrin et al., 1999). We found that glutaraldehyde fixation consistently produced better cells for DAPI staining (both wild type and M6). All the results presented in Figs 2 and 3 and in Table 1 are from glutaraldehyde-fixed cells. Almost all cells strongly autofluoresce and, in the wild type, there is a correlation between the cells that have no DNA and the cells that do not autofluoresce (indicative of ‘ghost’ or lysed cells).
Table 1. . Image analysis data for M6 and wild-type M. voltae cells.
Mean cell diameter (µm)
Diameter of largest cell observed (µm)
No. of anucleate cells
No. of cells >4 µm
Wild type (n = 1339)
1.93 ± 0.19
M6 (n = 2211)
2.12 ± 0.71
In the M6 mutant, we observed a high percentage (≈ 20%) of cells lacking any DNA. Anucleate cells autofluoresce brightly, indicating intact cells with no cell lysis (Fig. 2). We also observed some M6 cells with only small amounts of DNA present in the nucleoid (Fig. 2), but with bright autofluorescence, suggesting a partial segregation defect or ‘guillotine’ effect.
Although the majority of the M6 cells appeared to be similar in size to the wild-type cells, 1–2% of cells were strikingly different – much larger – (≈ three- to fourfold the diameter of wild-type cells) with a pronounced lobed structure (Figs 2C and 3). We call these titan cells. We observed titan cells in all M6 cultures examined. The enlarged mutants show large amounts of chromosomal DNA and retain the ability to autofluoresce (Fig. 2C). There are unusual darkened spots inside the enlarged cells (arrows in Fig. 3), which autofluoresce brightly.
Image analysis of thousands of cells from four different culture preparations, summarized in Table 1, clearly demonstrated the anucleate and titan cell phenotype. Wild-type cells produced a mean cell diameter of 1.9 ± 0.19 µm (n = 1339), and the largest observed wild-type cell was 3.75 µm (shown in Fig. 2A), while the M6 culture (n = 2211) had a mean cell diameter of 2.12 µm ± 0.71, and the largest observed cell was 9.1 µm. M6 cells clearly have a much larger fraction of cells with little or no DNA.
Defects in cell growth of smc::pac mutants
We expected a temperature-sensitive phenotype, as work in other prokaryotes had found that smc deletions (as well as mukB mutations) conferred a temperature-sensitive phenotype (Britton et al., 1998; Moriya et al., 1998; Jensen and Shapiro, 1999). At 30°C, with no agitation, we observed growth at 65% (± 12%) of wild-type growth. Growth at 37°C and 40°C shows decreasing growth rates relative to the wild type; this is even more pronounced in agitated cultures. Growth was completely inhibited above 41°C. The optimum growing temperature for wild-type M. voltae is 38°C (Whitman et al., 1982). These results are summarized in Table 2.
Table 2. Defects in growth of M. voltae M6.
% of wild-type growth
However, if M6 cultures are incubated at the non-permissive (43°C) temperature for 1 week, they remain viable when shifted to permissive temperatures (30°C), indicating that the temperature sensitivity is bacteriostatic. Also, M6 cultures were able to grow at 43°C in the absence of puromycin and, after growth in selection-free media, such cultures were still capable of growth in puromycin and still had the large cell and anucleate phenotype. This led us to question whether the temperature-sensitive phenotype resulted from the insertion or the puromycin resistance gene itself. As a control for phenotypic effects of smc insertion, we obtained M. voltae strain PS flaB2::pac. This strain has the same pac/pUC vector inserted into the flagellin gene flaB2 (Jarrell et al., 1996). This strain was also incapable of growth above 40°C and showed reduced growth relative to the wild type at 37°C. This flaB2::pac was also able to grow at 43°C in the absence of puromycin and, after growth in selection-free media, was still capable of growth in puromycin.
Thus, although we did observe a temperature-sensitive phenotype, we believe this to be due to the pac gene conferring puromycin resistance and not smc disruption.
Restoration of wild-type phenotype after reintroduction of the smc gene
We were able to restore wild-type nucleoid and cell size phenotypes by transforming M6 strains with a plasmid similar to pSL-1 but having the entire 3′ end of the smc gene next to the pac resistance cassette. We reasoned that serial subculturing in the early exponential phase will enrich for cells not producing anucleate daughter cells. After seven subcultures, we plated the cells and screened 100 colonies for the presence of a wild-type smc gene. We found two isolates that had an intact smc followed by the puromycin resistance cassette. These isolates produced a 2.8 kbp hybridizing band when EcoRI-digested DNA was probed with the SMC-I probe, whereas the M6 DNA produced three bands of ≈ 0.5, 1.5 and 1.8 kbp, and the wild-type DNA produced a single band of ≈ 2.9 kbp. These two isolates did not have an additional interrupted smc as probing with both the SMC-N probe and the SMC-I probe produced only a single band (results not shown). We observed the nucleoid and cell morphology (as for wild type and M6) and could discern no difference from wild-type cells – the phenotypes had been reversed. However, both isolates, although growing better than M6, still displayed growth defects relative to the wild type (results not shown).
Cell division in archaea, as prokaryotes with some eukaryotic cell division genes, has intrigued many commentators (Edgell and Doolittle, 1997; Bernander, 1998; 2003; Faguy and Doolittle, 1998; Tye, 2000; Kelman and Kelman, 2003). In this study, we sought to use genetics to illuminate the role that SMC plays in archaeal cell division. The objective of this study was to create an insertion into the smc gene in the archaeon M. voltae. This region is essential for a fully functional SMC in other organisms (Koshland and Strunnikov, 1996). We chose smc as a target for three reasons: (i) it is one of the few cell division genes that have viable knock-out phenotypes in other organisms; (ii) it plays a vital role in chromosome segregation linking DNA replication and cytokinesis; and (iii) it is one of the few cell division genes present in all three domains of life. The insertion in M6 deletes the entire C-terminal domain, which has been shown in other organisms to be important for DNA binding (Akhmedov et al., 1998; 1999; Jessberger, 2002). We expected an inactivation of smc to have some or all the following phenotypes: anucleate cells, decondensed chromosomes and temperature-sensitive growth. We observed ≈ 20% of the cells lacking DNA and a striking giant cell phenotype in 2% of cells.
We are confident that the anucleate and large cell phenotypes are the result of the lack of a full-length functional SMC protein. Southern blots clearly demonstrate an insertion into the smc gene. The likelihood of double insertions is small considering the relative rarity of single insertions of transformants and that, in the integration plasmid pSL-I, there were no other regions of extensive homology for recombination with M. voltae genomic DNA. Also, Southern blots using the SMC-I probe should have detected additional bands if the entire plasmid had inserted elsewhere. We do not believe that the phenotype results from polar effects, as it appears from sequence analysis that smc is transcribed as a single gene in M. voltae.
We had initially attempted to create a 5′ insertion in the smc gene, thus creating a true null mutant, but were unsuccessful. There is insufficient evidence to hypothesize whether this resulted from a lethal phenotype (as occurs with Δsmc in eukaryotes and in B. subtilis grown in complex media) or our particular insertion construct, selection or transformation procedure.
It is possible, however, that the observed phenotypes are caused by more complex effects than the insertional inactivation. For example, the observed effects could be due to the presence of a truncated SMC protein, rather than the absence of a full-length SMC. Ideally, one would like to observe reversal of the phenotype after complementation by a wild-type copy of the gene. Unfortunately, genetic systems that would allow expression of the wild-type gene from a plasmid are not available in M. voltae. We have used two controls, however, to ensure that the observed effects are the result of the smc locus and not the expression of the puromycin transacetylase gene, or any other gene, on the insertion plasmid. First, we obtained another puromycin-resistant insertion mutant (flaB2::pac) of M. voltae, which did not display any phenotypic difference in cell or nucleoid morphology (results not shown). To our surprise, the flaB2::pac mutant did show a defect in cell growth at elevated temperatures, indicating that puromycin resistance might be causing that phenotype. Secondly, we constructed a plasmid containing the 3′ end of the M. voltae smc gene next to the puromycin cassette. After transforming strain M6, the wild-type smc gene was restored (as seen by Southern blot), with the puromycin resistance gene beginning 48 bp downstream from the stop codon of smc. This construct also did not display any phenotypic difference in cell or nucleoid morphology with respect to the wild type. Thus, by restoring the wild-type smc gene, we can restore the wild-type cell and nucleoid phenotype, even in with puromycin selection. Taken together with what we know of smc function in other organisms, this work provides very strong evidence that the smc gene in M. voltae functions in chromosome segregation.
Anucleate cells and nucleoid morphology
We found 20% anucleate cells in the M6 culture. This is a higher fraction of anucleate cells than is observed in B. subtilis Δsmc or in E. coli ΔmukB. Although we also observed a few wild-type cells without high levels of DAPI stain, almost all had significant DAPI staining (Fig. 2). M6 cultures (except where noted below) consistently produced a large fraction of anucleate cells. We did not observe any apparent increase in the numbers of anucleate cells or in the presence of enlarged cells under shaking conditions or under higher growth temperatures. The anucleate cells autofluoresced brightly (Fig. 2), making it unlikely that the smc defect caused apparent anucleate cells by increasing the sensitivity of cells to lysis.
Although we thought that there might be some effect on cell morphology, we were surprised to observe the formation of enlarged cells (titan cells) with up to 20 times the volume of the average wild-type cell. The DNA content of these large cells is much greater than the average DNA content of M6 cells or wild-type cells. Several large M6 cells had 10–20 times the DAPI staining intensity compared with the average wild-type cell. The largest DAPI intensity in wild-type cells, as determined by ROI image analysis, was 87 125; in M6 cells, the largest DAPI intensity was 444 871. This is indicative of a large increase in DNA content per cell for titan cells. The cell cycle state of these titan cells is unknown, and methods are not available to determine cell cycle state or to synchronize M. voltae but, based on the size, DNA content and nucleoid structure. we would speculate that they are arrested precytokinesis, post replication, with replication only loosely tied to other cell cycle events. There is some supporting evidence for such a supposition. Malandrin et al. (1999) showed that, in M. jannaschii, a close relative of M. voltae, the DNA content per cell was unexpectedly high in both exponential and stationary phase cells. They estimated that exponential phase cells had 3–15 genome equivalents per cell. They also demonstrated, in contrast to bacterial systems, asynchronous nucleoid distribution (either from asynchronous segregation or asynchronous origin firing). Finally, as in all archaea studied to date but unlike all bacteria, they found > 1 genome equivalent in stationary phase cells. Also, Hjort and Bernander (2001) found in the distantly related archaeon Sulfolobus solfataricus that reinitiation of replication requires chromosome segregation but not cytokinesis. All of this tends to suggest that M. voltae does not tightly couple replication initiation to later events in the cell cycle (i.e. chromosome segregation and cytokinesis).
Slightly enlarged cells have been observed in B. subtilis Δsmc cells (Moriya et al., 1998) and in E. coli ΔmukB (Niki et al., 1991) when grown at the non-permissive temperature, but were not observed in C. crescentus Δsmc (Jensen and Shapiro, 1999). This titan cell phenotype in M. voltae is, perhaps, more reminiscent of the ttn (titan) seed phenotype in Arabidopsis (Liu and Meinke, 1998), which includes giant cells, polyploid nuclei and abnormal mitoses. These ttn mutations map to several smc and smc-associated genes in Arabidopsis (Liu et al., 2002).
Although C. crescentus Δsmc cells did not appear to be enlarged, the synchronized cell cycle of this organism allowed determination of the cell cycle state of mutant cells. While similar to wild-type cells at the permissive temperature, at the non-permissive temperature, a high proportion of the C. crescentus Δsmc culture accumulated at the predivisional cell stage, indicative of a cell cycle checkpoint (Jensen and Shapiro, 1999). This also explains the lack of anucleate cells as they are arrested before entering cytokinesis. The titan phenotype of Arabidopsis has also been linked to a cell cycle checkpoint (Liu et al., 2002), as has the sporulation defect of B. subtilis Δsmc mutants (Britton et al., 1998; Burkholder et al., 2001). Checkpoint mechanisms that block cell cycle progression or development in response to DNA damage or defects in replication are well described in bacteria and eukaryotes (Hartwell, 1992; Newton and Ohta, 1992; Autret et al., 1997; Burkholder et al., 2001). Work in the crenarcheote S. solfataricus has demonstrated cell cycle checkpoints in archaea (Jansson et al., 2000; Hjort and Bernander, 2001), but there are no details on the molecular mechanisms involved.
While titan cells contained large amounts of DNA, in more than 20 observations, they contained a single extended nucleoid (Figs 2C and 3), indicating that chromosome segregation is defective. The largest wild-type cells contained two segregated nucleoids (Fig. 2A). We frequently observed phase-dark foci on titan cells (Fig. 3), but they were found only rarely on wild-type cells (but were observed on several wild-type cells).
Archaea: bacterial cytokinetic machinery but eukaryotic replication machinery
The significance of the work lies not only in the results presented here, but more generally in using the power of genetics to investigate archaeal cell division mechanisms. We are currently investigating other potential phenotypes to M6 (e.g. DNA repair defects). Bernander (2003), in a recent review, characterized the archaeal chromosome segregation machinery as largely unidentified and uncharacterized. This study helps to begin the identification and characterization of archaeal chromosome segregation.
Microorganism strains and media
Methanococcus voltae strain PS (ATCC 33273; obtained from Professor Ken Jarrell, Queen's University, Kingston, Canada) was grown anaerobically in methanogenium medium DSMZ 141 at 37°C (except where noted) under an atmosphere of CO2/H2 (20:80) as described previously (Sowers and Noll, 1995). M. voltae mutant strain M6 (smc6::pac/pUC) was grown in methanogenium medium DSMZ 141 at 30°C under an atmosphere of CO2/H2 (20:80) and supplemented with 7.5 µg ml−1 puromycin. TOP10 chemically competent E. coli (Invitrogen) was grown in Luria–Bertani (LB) medium and supplemented with 50 µg ml−1 ampicillin when required. The strains, plasmids, probes and primers used are summarized in Table 3.
pac/pUC integration vector containing internal smc fragment produced by primers IP-F and IP-R
Primer or probe
Italicized nucleotides of primers represent added terminal restriction sites.
IP-F: corresponding to smc nucleotides 1847–1860
IP-R: corresponding to smc nucleotides 2347–2333
NTP-F: corresponding to smc nucleotides 101–116
NTP-R: corresponding to smc nucleotides 627–608
CT-F: corresponding to smc nucleotides 2054–69
CT-R corresponding to nucleotides 48–33
3′ to the smc stop codon
507 bp DIG-labelled PCR products of primers NTP-F and NTP-R
500 bp DIG-labelled PCR product matching the internal smc fragment produced by primers IP-F and IP-R in pSLI-1
Determination of smc gene sequence
The complete smc gene sequence was obtained from the M. voltae genome sequencing project (R. Feldman, R. Overbeek and W. B. Whitman, personal communication). The smc gene and the surrounding region have been deposited in GenBank as accession no. AY288521 (R. Feldman, R. Overbeek and W. B. Whitman, personal communication).
Construction of insertion vector
To construct a vector for inactivation of the SMC C-terminal region, a 500 bp internal fragment of the smc gene was polymerase chain reaction (PCR) amplified using primers IP-F and IP-R. The cycling conditions were as follows: 95°C for 5 min; 28 cycles of 92°C for 1 min, 59°C for 30 s, 72°C for 1min; and a final extension step at 72°C for 5 min. Genomic DNA isolated from wild-type M. voltae using a previously described method (Kalmokoff and Jarrell, 1991; Jarrell et al., 1996) served as a template for PCR. The internal fragment produced corresponds to nucleotide positions 1747–2247 of M. voltae smc gene (GenBank AY288521). 5′BamHI and 3′EcoRI terminal restriction sites (italicized on primers) were incorporated into the ends of the primers. The internal smc fragment was cloned into pCR4-TOPO using the TOPO TA cloning kit (Invitrogen), digested with EcoRI and purified electrophoretically using Ultrafree-DA (Millipore). After purification, the EcoRI-digested insert was cloned into the EcoRI site of a dephosphorylated pac/pUC integration vector (obtained from Professor Ken Jarrell) developed for transformation of M. voltae (Possot et al., 1988; Gernhardt et al., 1990; Berghofer and Klein, 1995) and designated pSLI-1. The pac expression unit in the vector contains the puromycin transacetylase gene of Streptomyces alboniger and is under transcriptional control of the M. voltae mcr promoter and terminator (Gernhardt et al., 1990). The resulting construct, pSLI-1, was transformed into chemically competent E. coli cells, purified by a QIAprep spin miniprep kit (Qiagen) and the orientation verified by standard dideoxy sequencing method. The smc insert is in the same orientation as the mcr promoter in the vector.
For restoration of the wild-type smc gene, we constructed an insertion vector containing the entire C-terminus of the M. voltae wild-type smc gene. Briefly, primers CT-F and CT-R were used to amplify (as above) a 1592 bp fragment from nucleotide 2054 of the M. voltae smc gene to 48 bp after the stop codon of the smc gene. This fragment was cloned into pac/pUC as above.
Transformation of M. voltae
For transformation of M. voltae, a liposome delivery method described previously (Metcalf et al., 1997; Thomas et al., 2001) was used with some modifications to the protocol. All steps in the procedure were performed in anaerobic conditions, and all buffers and solutions were prereduced by flushing with anaerobic gas. Liposomes containing vector DNA were prepared by combining 3 µg of the plasmid pSLI-1 (in a total of 22.5 µl of 20 mM Hepes buffer, pH 7.4) with 2.5 µl of Escort transfection reagent (Sigma) for 10 min. Samples of 6 ml of cells grown overnight with gentle shaking were aliquoted into 1.5 ml tubes and centrifuged for 5 min in a Galaxy mini microcentrifuge (VWR) placed inside an anaerobic chamber containing 5% CO2, 10% H2 and 85% N2. The supernatant was removed, and the cell pellet was resuspended in 0.75 ml of 0.85 M sucrose. Protoplasted cells (500 µl) were then added to the DNA–liposome mixture and incubated anaerobically at room temperature for 1.5 h. Control transformations of protoplasted cells with liposomes lacking plasmid DNA were also performed. The cells were then added to 10 ml of DSMZ 141 medium (without selection) and incubated overnight at 30°C without agitation. The 10 ml overnight cultures were centrifuged, and the supernatant was removed and resuspended in 500 µl of DSMZ 141 medium. Aliquots of 100 µl of cells were then spread on bottle plates containing DSMZ 141 medium and 1.5% (w/v) agar and supplemented with 7.5 µg ml−1 puromycin to select for puromycin-resistant transformants. After spreading of the transformants, the bottle plates were removed from the anaerobic chamber, and the headspace was filled with CO2/H2 (20:80). The bottle plates were then incubated at 30°C, and isolated colonies were visible in 15–20 days. Isolated colonies were then picked anaerobically and inoculated into DSMZ 141 media containing 7.5 µg ml−1 puromycin.
For restoration of the wild-type gene, we transformed M. voltae M6 (smc::pac) as above for wild-type M. voltae. After recovery overnight in the absence of selection, the 10 µl cultures were inoculated into 1 ml cultures containing 7.5 µg ml−1 puromycin and incubated at 30°C for 72 h. This culture was then subcultured every 72 h for a total of five subcultures, after which the liquid cultures were plated on puromycin plates. We screened 100 colonies by Southern analysis and found two that had a complete intact smc followed by the puromycin resistance cassette. These isolates had only one copy (as determined by Southern blots) of smc.
Southern blot hybridization analyses
Genomic DNA of both wild-type and mutant strains of M. voltae was isolated and digested with restriction enzymes, electrophoresed and transferred to nylon Biodyne A membranes (Pall) by capillary transfer. Two digoxigenin (DIG)-labelled smc DNA probes were produced by amplifying fragments of the smc gene using DIG-dUTP (Roche Molecular Biochemicals) by PCR, as directed by the manufacturer. Genomic wild-type M. voltae DNA was used as a template for the reaction. The primers and probes produced are listed in Table 3. Hybridizations were performed according to the manufacturer's instructions.
To examine cell and nucleoid morphology, wild-type and mutant cells were grown in liquid culture to mid-exponential phase (OD600≈ 0.4–0.6) and fixed by the addition of 25% glutaraldehyde to a final concentration of 3% (v/v) and incubated on ice for 20 min. As an alternative fixation, cells were harvested and resuspended in M. voltae salts (DSMZ medium 141 without yeast extract, peptone, vitamins), then fixed by the addition of 100% methanol at −20°C. Cells were placed on ice for 20 min. The cells were then centrifuged at 3000 g for 5 min, the supernatant was removed and the cells were washed once in M. voltae salts, then resuspended in M. voltae salts supplemented with 1.5 µg ml−1 DAPI (4′,6-diamidino-2-phenylindole; Molecular Probes). Cells were mounted on agarose-coated slides (100 µl of a 1% agarose solution was allowed to dry on to slides) and examined in a Nikon E-600 microscope with a mercury lamp. Images were obtained with a Coolsnap CCD camera (Roper Scientific). Separate images were taken under phase-contrast illumination, under fluorescent illumination with filter set UV-2E/C (Nikon) for DAPI staining and under fluorescent illumination with filter set Sapphire/UV GFP (Nikon) for autofluorescent images. Statistics were obtained from images by ROI analysis using Kodak 1D version 3.6.1 image analysis software. Composite images were created by layering in Adobe photoshop, version 7.0.
Growth rate determination
Cultures were inoculated into liquid media under an atmosphere of CO2/H2 (20:80) and grown at 30°C to mid-exponential phase. These 30°C cultures were used to inoculate 20 ml bottles of liquid media (supplemented with puromycin where appropriate) and incubated at 30°C, 37°C, 40°C or 43°C. Growth rates were recorded using OD600 values from an Eppendorf BioPhotometer.
The authors would like to Ken Jarrell and Sonia Bardy for technical help with M. voltae transformation, Ken Jarrell for the gifts of the pac/pUC vector and M. voltae PS culture, Julia Sanchez for assistance with microscopy, and R. Feldman, R. Overbeek and W. B. Whitman for providing the smc gene sequence before publication of the M. voltae genome sequence. Finally, we thank Kate Faguy, Ken Jarrell, W. B. Whitman, Sam Loker and Zvi Kelman for invaluable comments on the manuscript. The microscope used in this work was purchased with support from Army Research Office grant no. DAAD19-00-1-0538 to D.M.F.