INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

DNA of prokaryotic and eukaryotic cells is usually modified by methylation. This modification is carried out by DNA-specific methyltransferases, which transfer methyl groups from the universal substrate S-adenosylmethionine (AdoMet) to specific target sequences in the host DNA. On the basis of different chemical reactions, methyltransferases can be divided into three different groups leading to the appearance of N⁶-methyladenosine, C⁵-methylcytosine, or C⁴-methylcytosine. Amino acid sequence alignments of enzymes from these groups revealed conserved motifs characteristic of each methyltransferase family (25). In the enteric bacterium Escherichia coli, three methyltransferases are present: (i) Dam (DNA adenine methyltransferase), creating N⁶-methyladenosine in the specific target sequence 5'-G^mATC-3'; (ii) Dcm (DNA cytosine methyltransferase), leading to an internal C⁵-methylcytosine in the specific target sequence 5'-C^mC(A/T)GG-3'; and (iii) methyltransferase M · EcoK, modifying the second adenine in the sequence A^mAC(N₆)GTCG, which is specifically recognized as an unmethylated sequence by restriction endonuclease EcoK, which is part of the host-specific restriction-modification (R-M) system (for a review, see reference 22). Methylation of cytosine in CG dinucleotides dominates in cells of higher eukaryotes. In these cells, the degree of methylation is tightly controlled and plays an important role in the regulation of gene expression, organization of chromatin, and developmental and DNA repair processes (4).

Methyltransferases involved in R-M systems have been found in all taxonomic groups of eubacteria examined so far. R-M systems typically comprise two enzymatic activities, a restriction endonuclease and a methyltransferase. The genes of both enzymes are often organized in one operon. By extensive screening, more than 2,750 type II restriction endonucleases with a total of 211 different specificities have been found (27). Cyanobacteria represent a rich source of restriction enzymes. In several strains, two or three different enzymes are present; among them, isoschizomers of the enzyme AvaII have been often found (10). Six restriction endonucleases with different specificities were detected in the cyanobacterium Dactylococcopsis salina (19). The cognate methyltransferases have been investigated less. The genes encoding AvaI and AquI methyltransferases were cloned and sequenced from Anabaena variabilis and Agmenellum quadruplicatum, respectively. The amino acid sequences of both proteins show significant similarities to those of functionally related enzymes from heterotrophic bacteria (13, 30).

Genetic tools are well developed only for a few cyanobacterial model strains. A general problem in establishing such systems for other strains is the barriers made by endogenous restriction systems that prevent stable access of foreign DNA into the cells. As has been found before in many instances, in vitro methylation of DNA also improves the efficiency of DNA transfer into cyanobacterial strains (35). The cyanobacterium Synechocystis sp. strain PCC 6803 (hereafter referred to as Synechocystis) belongs to the group of naturally competent strains which are transformable by free exogenous DNA (6). A Ca²⁺-dependent nuclease located in the cytoplasmic membrane is involved in DNA uptake and is assumed to convert double-stranded DNA into the single-stranded form (3). Recently, the entire genome of Synechocystis was sequenced, making this strain the favored cyanobacterium for genetic studies (11).

In this work, we examined the chromosomal DNA of Synechocystis for modifications by methylation. Sequence-specific methylations were detectable, whereas no restriction endonuclease activities were found. The open reading frame (ORF) slr0214 (11), which shows high similarities to genes encoding DNA-specific methyltransferases, was inactivated on the chromosome. In the mutant, alterations of the DNA methylation pattern were detected, demonstrating that slr0214 encodes a methyltransferase which recognizes a target sequence of 6 bp. Significantly, compared to the wild-type (WT) strain, the mutant could not grow in CO₂-enriched cultures under standard conditions, indicating an important function of this methyltransferase in Synechocystis. Furthermore, the slr0214 gene was subcloned and expressed in E. coli. The resulting protein showed methyltransferase activity in vitro.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains and culture conditions. A derivative of Synechocystis with enhanced transforming capacity was used in all experiments and was obtained from S. Shestakov (Moscow State University, Moscow, Russia). Axenic cells were cultured on agar plates at 30°C under constant illumination with mineral medium C (16). Transformants were initially selected on medium containing 10 µg of kanamycin (Sigma) ml¹, while the segregation of clones and the cultivation of mutants were performed with 50 µg of kanamycin ml¹. For physiological characterization, axenic cultures of cyanobacteria (about 10⁸ cells ml¹) were grown photoautotrophically under continuous illumination with 180 µmol of photons s¹ m² (warm light; Osram L58 W32/3) at 29°C in batch cultures bubbling with air enriched with CO₂ (2.5%) in potassium nitrate-containing growth medium (1). Contamination by heterotrophic bacteria was checked by microscopy or spreading of 0.2 ml of culture on Luria-Bertani (LB) plates. E. coli JM101 (31) was used for routine DNA manipulations. For the overexpression of protein, protease-deficient E. coli BL21 (31) was used. E. coli was cultured in LB medium at 37°C. Growth was monitored by measuring the optical density at 750 nm (OD₇₅₀) for Synechocystis and at 500 nm (OD₅₀₀) for E. coli.

DNA manipulations. The isolation of total DNA from Synechocystis was done as described previously (2). All other DNA techniques, such as plasmid isolation, transformation of E. coli, ligation, and restriction analysis (restriction enzymes were obtained from Life Technologies and New England BioLabs), were standard methods (31). For Southern hybridization experiments, the DNA probes were labelled with digoxigenin by use of a PCR DIG probe synthesis kit (Boehringer Mannheim Biochemicals). For restriction analyses with chromosomal DNA from Synechocystis, restriction endonucleases were used in a 10-fold excess and were incubated for at least 16 h at 37°C in order to ensure complete digestion. The following synthetic primers were specifically used for amplification of the sequence encoding the DNA-specific methyltransferase (slr0214) (11): 5'-GGCGCCGGATCCATGGCCAGACCCATTGCCAT-3' (Met-fw; BamHI recognition sequence underlined; start codon of slr0214 in boldface) and 5'-CGGCGCGAATTCTTAGGAATGGGATTTGGAC-3' (Met-re; EcoRI recognition sequence underlined; stop codon of slr0214 in boldface). In order to generate an slr0214 deletion mutant, a fragment containing slr0214 and two neighboring ORFs was amplified with the following primers: 5'-CGCGGATCCCCGAGTTTTATTAGCCCTTT-3' (Met-fw; BamHI recognition sequence underlined) and 5'-AAAACTGCAGCTTTAGTCTCAGTGTGGCGA-3' (Met-re; PstI recognition sequence underlined). For the in vitro methylation assay, an internal fragment of the clpC gene of Synechocystis was amplified by PCR with the following primers: 5'-CTCAAATCCATGGGGGTTAA-3' (Clpfw) and 5'-GAGGTCATGATCAACAGGGT-3' (Clpcre) (custom oligonucleotide synthesis; Pharmacia). For PCR, PCR-supermix (Life Technologies) and the following temperature cycles (30 times) were applied: 15 s at 94°C, 30 s at 52°C, and 1 min at 72°C. The plasmid vectors pUC19 (38) and pGEM7 (Promega) were used.

Generation of insertion and deletion mutants. For the generation of slr0214 mutants, the aphII gene cartridge (encoding aminoglycoside phosphotransferase II) from E. coli plasmid pUC4K (37) was integrated at selected restriction sites into the sequence of slr0214 cloned into E. coli vectors (Table 1). Plasmid DNA of these constructs was isolated from E. coli with a QIAprep spin plasmid mini kit (Qiagen). About 1 µg of DNA was used for the transformation of Synechocystis, and kanamycin-resistant clones were selected (9).

Protein overexpression. For overexpression and purification of the Slr0214 protein, the glutathione S-transferase (GST) gene fusion system (Pharmacia) was used. The slr0214 ORF was amplified from chromosomal DNA of Synechocystis by PCR with primers Met-fw and Met-re. The translational start codon (boldface letters in the Met-fw primer) was located immediately behind a BamHI site used to clone the fragment in frame with the GST ORF into pGEX-4T-3 (Pharmacia). Constructs showing the correct insertion were transformed into E. coli BL21, and a transformant was selected. The cells were cultured at 30°C in LB medium. The expression of the protein was induced at an OD₅₀₀ of 1.0 by the addition of isopropyl--D-thiogalactopyranoside (IPTG) (100 µM), and incubation was continued for 150 min. The proteins were extracted from E. coli by sonication, and the fusion protein was bound to a glutathione-Sepharose slurry (Pharmacia). The Slr0214 protein was released from the matrix by cleavage of the protease thrombin (Boehringer), for which a cleavage site was inserted between GST and Slr0214. The extracted proteins were separated in sodium dodecyl sulfate gels containing acrylamide gradients from 7.5 to 15% in the buffer system of Laemmli (17).

Protein extraction and methyltransferase assay. The extraction of proteins from frozen cell material of Synechocystis was done by ultrasonic treatment (30 W, 2 min) in 10 mM HEPES-NaOH buffer (pH 7.3) under ice cooling. Crude protein extracts were purified by centrifugation (15 min, 5,000 × g, 4°C). The protein content was estimated by the method of Lowry et al. (20). DNA-specific methyltransferase activity was assayed by incubation of unmethylated chromosomal DNA of Micrococcus lysodeikticus (Sigma) with [³H]AdoMet (Amersham) as described by Günthert et al. (7). The radioactivity incorporated in the DNA was determined by liquid scintillation counting. In vitro methylation of the internal fragment of the clpC gene was performed after mixing about 1 µg of DNA, AdoMet (10 mM final concentration), and low-salt restriction buffer (Life Technologies) in a total volume of 200 µl. The reaction was started by adding about 0.5 µg of the Slr0214 protein, and the mixture was incubated for 12 h at 37°C. After concentration by precipitation, the DNA was directly used for restriction analysis.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

The occurrence of sequence-specific DNA methylation was examined by incubation of chromosomal DNA isolated from Synechocystis with different restriction enzymes to test whether or not their activities are specifically influenced by methylation of their recognition sequences. In these experiments, it was found that the DNA was resistant to the action of MboI, ApaI, EaeI, HaeIII, PvuI, and SgfI, while other restriction endonucleases were able to cut (Table 2). Different sensitivities of the DNA to different restriction endonucleases indicated that at least three methylation activities seemed to be present. One represented a Dam-like methyltransferase, which was demonstrated by the inhibition of the restriction enzyme MboI and the activities of DpnI and Sau3A (Table 2). The activities of these enzymes were differentially influenced by methylation of the adenine in their recognition sequence, 5'-GATC-3': MboI could not cleave Dam-modified DNA; DpnI could cleave only Dam-modified DNA; and Sau3A was not influenced by Dam modification, since its cognate methyltransferase modifies the cytosine (24). Comparison of the recognition sequences of the other inhibited restriction endonucleases showed that at least two different cytosine-specific methyltransferases were involved. The activities of ApaI, EaeI, and HaeIII were most probably affected by one enzyme recognizing the sequence 5'-GGCC-3'. PvuI and SgfI should have been inhibited by a cytosine-specific methyltransferase modifying one of the two cytosines within the sequence 5'-CGATCG-3', which includes the Dam recognition site. The complete digestion of chromosomal DNA by Sau3A excluded methylation of the second cytosine. Thus, this cytosine-specific methyltransferase most probably modified the first cytosine of a PvuI site (Table 2). Statistical analyses showed that the frequency of occurrence of PvuI and SgfI recognition sequences was increased severalfold in the chromosomal DNA of Synechocystis compared to their theoretical frequency in a random sequence. The frequencies of the recognition sequences of EaeI and HaeIII were also increased but to a lesser extent, whereas the occurrence of the MboI site corresponded to its frequency in a random sequence (Table 2).

In order to test if these methyltransferase activities were part of a strain-specific R-M system, the restriction endonuclease activity of Synechocystis was examined. This was done by incubating unmethylated DNA of phage  (MBI Fermentas) with protein extracts (150 µg) in the presence of the basic reaction buffers (React 1 to 4; Life Technologies) recommended for most restriction endonucleases at 37°C overnight (19). These experiments clearly showed that there was no restriction endonuclease activity, since the  DNA was not cut into defined fragments after incubation with the large amounts of proteins used over a long time under different reaction conditions (data not shown).

On the genome of Synechocystis, there is one ORF (slr0214) encoding a putative cytosine-specific methyltransferase (11). The amino acid sequence of Slr0214 was aligned with those of other methyltransferases. The proteins showing the best alignments all belonged to the group of C⁵-cytosine methyltransferases (Fig. 1). The highest amino acid identity (48.2%) over the whole sequence was found with the XorII methyltransferase from Xanthomonas oryzae (5). The homology pattern characteristic for the group of C⁵-cytosine methyltransferases was found to be well conserved in the predicted amino acid sequence for Synechocystis (Fig. 1). All of the 10 conserved regions (26) could be identified. The variable region between motifs XIII and IX has been shown to determine the specificity of DNA sequence recognition (14). Especially within the variable region the amino acid sequences of Slr0214 from Synechocystis and the XorII methyltransferase showed the highest homology (41.5% identical amino acid residues; less than 30% identity to those of other enzymes), indicating that these two enzymes also might have similarities in their recognition sequences. XorII is an isoschizomer of PvuI, whose recognition sequence was found to be modified (Table 2). Therefore, the protein encoded by slr0214 could be a candidate for a methyltransferase that modifies the PvuI recognition sequence (5'-CGATCG-3') in Synechocystis. Since XorII is also inhibited by Dam modification (5), the following experiments were performed with PvuI.

View larger version (84K): [in this window] [in a new window]   FIG. 1.   Amino acid sequence comparison between the Slr0214 protein (SynMI) (11) and the C5-methylcytosine-synthesizing methyltransferases XorII (5), DdeIM (34), NgoII (33), and HphIM (21). Uppercase letters and a shaded background indicate amino acids identical to those in SynMI. The conserved motifs characteristic for C5-cytosine methyltransferases (26) are indicated by Roman numerals.

In order to analyze the function of the putative C⁵-cytosine methyltransferase, mutants impaired in slr0214 were generated. After cloning of a PCR fragment containing Slr0214 (resulting in plasmid pGGM02) (Table 1), an aphII gene cartridge (conferring kanamycin resistance) whose transcription occurred opposite to the direction of transcription of slr0214 was inserted into the central HincII site (resulting in plasmid pGGMK29) (Fig. 2 and Table 1). Furthermore, a PCR fragment containing about 1.6 kb in addition to the sequence of slr0214 was cloned (resulting in plasmid pGGML05) (Table 1). In plasmid pGGML05, about 60% of the slr0214 sequence was deleted by cutting with BalI, and the deleted fragment was replaced by an aphII gene cartridge (resulting in plasmid pGGMDE12) (Fig. 2 and Table 1). Synechocystis was transformed with these constructs. Since these plasmids do not replicate in Synechocystis, continued selection on kanamycin demands the integration of the aphII gene into the chromosome. After cultivation for several generations at an increased concentration of the antibiotic, we isolated kanamycin-resistant clones. With this strategy, mutants impaired in slr0214 were obtained by insertion of the kanamycin resistance cassette (slr0214 or insertion mutant) or by replacement of the cassette (slr0214 or deletion mutant).

View larger version (22K): [in this window] [in a new window]   FIG. 2.   Schematic drawing showing the genetic organization, restriction map, and protein-encoding regions of the chromosomal region encoding Slr0214 (SynMI) in Synechocystis (11). (A and B) The insertion of the aphII gene in selected sites to generate mutants is shown. (A) BalI was used to obtain the slr0214 mutant. (B) HincII was used to obtain the slr0214 mutant. (C) Protein-encoding region. (D) The binding sites of the primers are represented by triangles under the arrows indicating the protein-encoding regions in panel C.

The lesion in the DNA of these mutants was characterized by Southern hybridization experiments (Fig. 3; compare to Fig. 2). In NcoI-digested DNA of the slr0214 mutant, one fragment of 4.5 kb was found with the slr0214 gene probe. This fragment was 1.2 kb larger than the signal detected in the WT DNA, consistent with the integration of the 1.2-kb aphII gene cartridge (Fig. 3A). One fragment of only 2.4 kb became visible by hybridization with the slr0214 mutant DNA. This size exactly matched the size of the fragment expected for the deletion mutant (Fig. 3A). After cutting of chromosomal DNA with HindIII, in the WT only one fragment of 2.5 kb was detectable with the slr0214 gene probe, while in the mutants two fragments (the aphII gene contains one HindIII site) of the expected sizes (2.2 and 1.7 kb for the slr0214 mutant and 2.0 and 1.0 kb for the slr0214 mutant [with the slr0214 gene probe, the 1.0-kb fragment of the deletion mutant was detectable only as a very faint band, since it shared only 30 bp with the DNA probe]) became visible (Fig. 3A). The same pattern was obtained for the DNA of the mutants when the aphII gene was used as a probe (Fig. 3B). Again, after digestion with NcoI, fragments of 4.5 and 2.4 kb were detected. After cutting with HindIII, the aphII gene probe hybridized with a 2.2-kb fragment and a 1.7-kb fragment of the slr0214 mutant DNA, whereas a 2.0-kb band and a 1.0-kb band were clearly visible with the slr0214 mutant DNA. No specific signal could be detected by hybridization with the WT DNA. These hybridization patterns indicated that the plasmid constructs used for the transformation of Synechocystis to obtain the mutants were correctly integrated by double crossing over and replaced completely the WT copy of the slr0214 gene.

View larger version (84K): [in this window] [in a new window]   FIG. 3.   Southern blot experiments for characterization of complete segregation of the slr0214 mutant and the slr0214 mutant. The digoxigenin-labelled PCR fragment containing slr0214 (A) or the digoxigenin-labelled aphII gene (B) was used as a probe for hybridization to HindIII digested (lanes 1 to 3) and NcoI-digested (lanes 4 to 6) chromosomal DNAs from cells of the WT (lanes 1 and 4), the slr0214 mutant (lanes 2 and 5), and the slr0214 mutant (lanes 3 and 6) of Synechocystis. M, fragment sizes of digoxigenin-labelled HindIII-digested  DNA.

In order to analyze changes in the methylation pattern, chromosomal DNA isolated from the mutants was incubated with different restriction enzymes which were found to be inhibited by the WT DNA (Table 2). Digestion with NcoI was used as a control reaction in order to show that the DNA preparations could be cleaved under the selected reaction conditions. In contrast to the WT DNA, the DNA of the slr0214 mutant (Fig. 4) as well as the DNA of the slr0214 mutant (data not shown) became accessible to restriction by PvuI and SgfI. The other methylation activities, the Dam-like methylation and the second cytosine-specific methylation, remained intact, because the DNA of both mutants was still resistant to MboI as well as HaeIII and was digested by DpnI (Fig. 4; compare to Table 2). The ability of PvuI and SgfI to cut mutant DNA was confirmed by Southern hybridization experiments. With PvuI- and SgfI-cut DNAs from the WT and the mutants, three and two fragments (PvuI and SgfI, respectively) (Fig. 2) of the expected sizes appeared after application of the slr0214 gene probe, while with the WT DNA only hybridization to the uncut high-molecular-weight DNA was visible (data not shown). Therefore, the methyltransferase encoded by slr0214 is specifically directed to the recognition sequence of PvuI or its isoschizomer XorII. Furthermore, the methylation is directed to a cytosine within the recognition sequence, since the internal adenine remained modified, as shown by the inhibition of MboI. We propose the designation SynMI (Synechocystis methyltransferase I) for the slr0214-encoded enzyme. Beside this enzyme, a second cytosine-specific methyltransferase which is responsible for the modification of the HaeIII recognition sequence must exist in Synechocystis.

View larger version (94K): [in this window] [in a new window]   FIG. 4.   Separation of fragments generated during a restriction analysis of chromosomal DNAs of the WT (lanes 1, 3, 5, 7, 9, 11, 13, and 15) and the slr0214 mutant (lanes 2, 4, 6, 8, 10, 12, 14, and 16) of Synechocystis obtained by agarose gel electrophoresis. The following restriction endonucleases were applied in this experiment: lanes 1 and 2, uncut control (n. c.); lanes 3 and 4, SgfI; lanes 5 and 6, PvuI; lanes 7 and 8, NcoI; lanes 9 and 10, HaeIII; lanes 11 and 12, MboI; lanes 13 and 14, Sau3AI; and lanes 15 and 16, DpnI. M, fragment sizes of HindIII-digested  DNA.

In further experiments, slr0214 and slr0214 mutants of Synechocystis were characterized with regard to changes in their physiology. During cultivation on agar plates, no obvious differences between mutant and WT cells were observed. Surprisingly, the growth of the slr0214 mutant changed dramatically in CO₂-enriched cultures when the medium of Allen and Arnon (1) was used. It became very difficult or impossible to cultivate the mutant under the growth conditions described for the WT. It was possible to grow these cells only at higher densities or, alternatively, at lower light intensities. However, even at lower light intensities, the growth of the mutant was significantly reduced in comparison to that of the WT (Fig. 5). These changes in the growth characteristics implied that methyltransferase SynMI (Slr0214) seems to fulfill an important function in Synechocystis. Additionally, the total DNA-specific methyltransferase activity was measured in protein extracts obtained from cells of WT Synechocystis and the slr0214 mutant. The specific methyltransferase activity of this mutant was reduced to about 30% the WT activity (Fig. 5).

View larger version (23K): [in this window] [in a new window]   FIG. 5.   Comparison of the total methyltransferase activity in and the growth of cells of the WT (columns 1 and 3) and of the slr0214 mutant (columns 2 and 4) of Synechocystis after cultivation in CO2-gassed cultures with the medium of Allen and Arnon (1) at reduced light intensities (about 40 µmol of photons s1 m2). The data represent the mean values from two independent experiments (each done in duplicate).

The biochemical reaction of the SynMI (Slr0214) protein was confirmed by an in vitro assay. The slr0214 gene was cloned in pGEX4T3 and overexpressed in E. coli. From IPTG-induced E. coli cultures proteins were isolated and purified with a glutathione-Sepharose column. From the affinity matrix a protein of the expected size of 43 kDa was purified to homogeneity after cleavage by thrombin (Fig. 6A). The purified protein showed significant methyltransferase activity in an in vitro enzyme assay (Fig. 6B). The methyltransferase activity was detected only in the low-salt buffer; increased salt content in the buffer led to complete inhibition of the activity of SynMI. Furthermore, the methylation specificity of the purified SynMI protein was tested in vitro, confirming the PvuI recognition sequence as its target. A 1.8-kb internal fragment of the clpC gene of Synechocystis, which contains two PvuI sites, was amplified by PCR, methylated in vitro, and treated with restriction enzymes (Fig. 6C). While the unmethylated PCR fragment was completely cut by PvuI (fragments of 1.07 and 0.73 kb appeared), the methylated fragment was found to be fully resistant to the PvuI treatment. ClaI digestion was used as a control in order to show that the methylated fragment could be cut by enzymes with different specificities.

View larger version (47K): [in this window] [in a new window]   FIG. 6.   Overexpression of the Slr0214 protein in E. coli BL21 by use of the GST gene fusion system (Pharmacia). (A) Coomassie blue staining of proteins after separation by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Lanes: 1, total extract from E. coli cells after induction by IPTG; 2, crude extract after sonication and centrifugation at 25,000 × g for 20 min; 3, isolated Slr0214 protein after thrombin cleavage; M, prestained broad-range markers (Bio-Rad). (B) Methyltransferase activity of the purified Slr0214 protein in in vitro assays with buffers containing different NaCl concentrations. Lanes: 1, assays done with 100 mM NaCl; 2, assays done with 50 mM NaCl; 3, assays done under NaCl-free conditions. (C) Restriction analysis of a DNA fragment with (lanes 4 to 6) and without (lanes 1 to 3) methylation by SynMI in vitro. Lanes: 1 and 4, uncut; 2 and 5, PvuI; 3 and 6, ClaI; M, fragment sizes of EcoRI/HindIII-digested  DNA.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Extensive restriction analyses revealed that the chromosomal DNA of the cyanobacterium Synechocystis is modified by methylation. One DNA-specific methyltransferase activity was indicated by the resistance of chromosomal DNA to the restriction activity of PvuI and SgfI, which are known to be influenced by C⁵-cytosine methylation of their recognition sequences in CG dinucleotides (12, 32). The ORF slr0214 (11) was identified as the gene encoding the methyltransferase responsible for this modification. Three lines of evidence led to this conclusion. (i) By comparison of the Slr0214 amino acid sequence to related sequences, the highest degree of similarity was found to C⁵-cytosine methyltransferases. The N- and C-terminal parts are especially well conserved, which is characteristic for this group of enzymes, while the central part is variable and determines their sequence specificities (14). Among the related sequences, that of the XorII (an isoschizomer of PvuI) methyltransferase showed the highest degree of homology. (ii) After overexpression of the slr0214 gene in E. coli, it was possible to demonstrate methyltransferase activity in vitro. The methylation of PvuI recognition sequences was directly shown by use of methylated and unmethylated PCR fragments for restriction analysis. (iii) The final proof was obtained by analyzing the features of mutants impaired in slr0214; the insertion and deletion mutants showed the same alterations in phenotype. In both mutant types, the chromosomal DNA could be digested by PvuI as well as by SgfI, while in the WT, these sites were completely blocked. The PvuI site was identified as the recognition sequence, since in Southern blot experiments it became obvious that PvuI sites which were not part of an SgfI site were cut in the mutant DNA, while such PvuI sites were not cut in the WT DNA. Furthermore, the specific activity of methyltransferases was significantly decreased in extracts obtained from the mutants. Nevertheless, there was a significant residual activity, which can be explained by the occurrence of the other methyltransferases. The identity of the methylated base produced by SynMI in the PvuI recognition sequence has not yet been experimentally determined. However, it is likely that the first cytosine represents the target for methylation by SynMI, a conclusion which could be drawn from the data obtained for the activity and inhibition of different restriction enzymes affecting the PvuI site and its internal sequence, 5'-GATC-3', by use of DNAs from the WT and the Slr0214 mutant (Table 2 and Fig. 4).

At least a second type of methyltransferase activity is present in Synechocystis; this activity resembles the Dam modification characterized in E. coli, in which the adenine in the sequence 5'-GATC-3' is modified. This finding was clearly shown by the inhibition of MboI, which is affected by the adenine methylation, and by the activity of DpnI, which cuts only if the adenine is methylated (24). In E. coli, Dam is involved in multiple functions: (i) mismatch repair, (ii) regulation of initiation of chromosome replication, and (iii) modification of expression of several genes (for a review, see reference 22). Dam-like modification of chromosomal DNA had been found in several cyanobacteria (e.g., 18). In Synechocystis, mutation of slr0214 did not affect the Dam-like methylation type. The corresponding methyltransferase probably is encoded by ORF slr1803 (11), which encodes a putative adenine-specific methyltransferase in Synechocystis and shows the highest similarity (41.2% identical amino acid residues) to MboA, the cognate methyltransferase of the MboI restriction endonuclease in Moraxella bovis (36).

In addition, another cytosine-specific methyltransferase activity modifying the HaeIII recognition sequence was not affected by inactivation of slr0214. Except for slr0214 and slr1803, no other ORF encoding a putative DNA-specific methyltransferase has been identified on the sequenced chromosome of Synechocystis (11). Thus, the HaeIII site-specific methyltransferase might be encoded by one of the three extrachromosomal elements (with a total size of 280 kb) (15) which were identified in Synechocystis or, alternatively, this enzyme might represent a new type of methyltransferase. The latter is not very likely because of the high level of conservation of cytosine-specific methyltransferases (26).

In many bacteria, DNA methylation is related to a strain-specific R-M system. However, in Synechocystis, no restriction endonuclease activity was demonstrated by incubation of lambda DNA with large amounts of cellular protein. The capacity of this strain for natural transformation (6, 35) is a further indication for the absence of such enzymes; in addition, on the entire chromosome no ORF showed homologies to genes for known restriction endonucleases (11). Therefore, Synechocystis expresses methyltransferases which are not part of a host-specific R-M system. Nevertheless, at least the enzyme SynMI (Slr0214) seems to be have an important function, since the mutants showed altered growth characteristics.

The recognition sequence of the methyltransferase SynMI is part of the HIP1 sequence (highly iterated palindrome) (8), which is identical to the SgfI site. HIP1 is very abundant in the DNA of several but not all cyanobacteria, where it occurs in coding and intergenic regions (28). The complete genome sequence of Synechocystis (11) allowed us to estimate their frequencies. SynMI or PvuI sites were found every 698 bp and SgfI or HIP1 sequences were found every 1,131 bp in the DNA (Table 2). In Synechocystis sp. strain PCC 6301, HIP1 was found to be extremely frequent (every 320 bp). However, in this strain, HIP1 is not methylated at all or by the same methylase as in Synechocystis, since its chromosomal DNA was completely digested by PvuI (28). The HIP1 sequence seems to be involved in excision and gene rearrangement (29).

In Synechocystis, the target sequences of the SynMI methyltransferase, PvuI, including the HIP1 sequence, are highly overrepresented and are completely methylated in the WT. Hypothetically, these methylated sequences could serve as binding sites for chromatin-like proteins. The same function was suggested for another repetitive sequence found in filamentous cyanobacterial strains (23). The proteins associated with the DNA might have an influence on the rate of gene expression, as was found for histones in eukaryotic cells. Interestingly, the sequences of some highly expressed genes in the genome of Synechocystis (e.g., genes encoding for rRNAs and for subunits of photosystem I [psaF, -D, and -L] and of photosystem II [psbA1 to psbA3, psbCD]) (11) do not contain any PvuI site and are therefore also not methylated in WT cells. The complete absence of methylation in the slr0214 mutants of Synechocystis could disturb the balanced gene expression program, leading to problems in growth at higher light intensities in CO₂-enriched cultures. In future experiments, we will analyze the physiological function of the SynMI methyltransferase in more detail by growing the WT and the mutants under different conditions and by comparing the expression of genes with or without PvuI sites in both the WT and the mutants.