INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Cyanobacteria (blue-green algae) are an ancient group of prokaryotic microorganisms exhibiting the general characteristics of gram-negative bacteria. They are unique among the prokaryotes in possessing the capacity of oxygenic photosynthesis. In addition, some cyanobacteria also have the capacity for fixation of atmospheric nitrogen within the same organism. These qualities make cyanobacteria the most successful and widespread group among the prokaryotes found in diverse terrestrial and aquatic environments. In addition, some cyanobacteria form symbioses with an exceptionally broad range of representatives within the plant kingdom (reviewed in reference 28). These include plants from all divisions: Bryophyta (mosses, liverworts, and hornworts), Pteridophyta (aquatic ferns of the genus Azolla, approximately 7 species), gymnosperms of the family Cycadaceae (approximately 150 species), and angiosperms of the family Gunneraceae (approximately 50 species), as well as diverse lichenized fungi. Not only do cyanobacteria have a broad host range, but the infected symbiotic tissue also varies between the different plants. The cyanobacteria are found in extracellular cavities of the bryophyte thalli and of the Azolla leaves, extracellular in a zone of specialized roots of cycads, and intracellular in stem glands in Gunnera species (2, 28). The cyanobacteria comprise only a few percent of the host plant biomass, but in all interactions they are highly beneficial, making the host plants autotrophic with respect to nitrogen. With few exceptions, the cyanobacteria that enter into symbiosis belong to the filamentous heterocystous genus Nostoc (30). Isolation and the ability to grow the two partners separately under sterile conditions have made it possible to reconstitute the symbiosis. This has been demonstrated for the Anthoceros-Nostoc and Gunnera-Nostoc symbioses (4, 7, 13). In those experiments, it was shown that some Nostoc organisms isolated from one symbiotic association could form symbiosis with another host. Those results raise interesting questions concerning the specificity and diversity among symbiotic Nostoc, both within and between different host plants. However, very little is known about the diversity of symbiotic cyanobacteria. Studies based on restriction fragment length polymorphism (RFLP) and PCR techniques have recently been used to examine the Anabaena-Azolla symbiosis for classification of the cyanobacterial symbionts from different Azolla species (5, 8, 27). Although difficulties in culturing the symbiont on artificial medium have been a problem for such studies, it seems that little diversity exists among the symbionts from different Azolla species (5). Isolates from cycads and Gunnera have been studied with respect to genetic diversity by using protein profiles and the RFLP technique (19, 37, 38). Although the number of isolates included in those studies was limited, it could be concluded that diversity exists among the symbiotic isolates from different plant species. Moreover, among the isolates from different Gunnera species, the same hybridization pattern could be observed with glnA and nifH as probes, indicating that the isolates are similar or closely related (37).

Repetitive sequences constitute an important part of the prokaryotic genome. The repetitive extragenic palindromic (REP) (34) and enterobacterial repetitive intergenic consensus (ERIC) (11) sequences were originally described for the family Enterobacteriaceae but later found in several gram-negative bacteria and close relatives in the same phyla (35). For gram-positive bacteria, Martin et al. (21) described the BOX elements in Streptococcus pneumoniae. For cyanobacteria, a distinct family of repetitive sequences, the short tandemly repeated repetitive (STRR) sequences, have been described (12, 23). The STRR sequences have been identified in a number of cyanobacterial genera and species, all belonging to the heterocystous cyanobacteria (23). Initially the sequences were described for Calothrix species, where the copy number was estimated to about 100 per genome (23). In addition, a 37-bp long tandemly repeated repetitive (LTRR) sequence has recently been identified in Anabaena strain PCC 7120 (22). The repeated sequence was by hybridization experiments found to be present at a low copy number in Anabaena strain PCC 7120. The LTRR sequence was detected in both heterocystous and nonheterocystous cyanobacteria (22). Recently, the STRR sequence was demonstrated to be a valuable tool for identification and characterization of cyanobacteria. The STRR sequence was used as probe to identify toxin-producing cyanobacteria from a Finnish lake (32).

The function of repetitive sequences is still unclear. It has been suggested that they may regulate transcription termination (10) or be the target of DNA-binding proteins responsible for chromosomal maintenance in the cell (19, 23). However, the conserved status of these repetitive sequences makes them methodologically important tools for diversity studies among related prokaryotes and for identification (fingerprinting) of microorganisms in general.

The discovery of short repeated sequences dispersed in the genome of bacterial species formed the basis of a technique which utilizes oligonucleotide-derived repetitive sequences in the PCR, rep-PCR (35). The method has been used to fingerprint different bacteria (6, 15, 18, 31) and has been shown to be efficient for differentiating closely related strains (6). So far, the repetitive sequences identified in cyanobacteria, STRR and LTRR, have not been used to generate rep-PCR. Our objective was to develop an easy and reliable fingerprint method for cyanobacteria by using different oligonucleotides from repetitive sequences as primers in the PCR. An additional objective was to study the genetic diversity in a collection of cyanobacteria, with special emphasis on symbiotic isolates from the angiosperm Gunnera. The method developed could be demonstrated to be used directly on intact cyanobacterial filaments and cells.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cyanobacterial isolates and culture conditions. The cyanobacteria used in this study are listed in Table 1. Strains with a PCC number were obtained from the Collection Nationale de Cultures de Microorganismes, Institut Pasteur, Paris, France. The Gunnera-Nostoc isolates 8913, 8916, 8972, 8938, 8940, 8941, 8942, 8923, 8924, 8928, 8929, 8930, and 8964 were collected in 1988 and stored as dormant cultures in the dark on dry agar plates. Two months before use, they were reinitiated by transfer to liquid media. Other cyanobacteria listed in Table 1 were from continuously grown liquid cultures. All cyanobacteria used in this study were grown in BG-11 medium (33) at 28°C under continuous shaking and light as described by Johansson and Bergman (13). The following bacteria were included as controls: Escherichia coli, Rhizobium leguminosarum biovar trifolii, Pseudomonas fluorescens, and Bacillus megaterium. For cultivation of E. coli, LB medium (24) was used; TY medium (3) was used for growth of R. leguminosarum biovar trifolii, P. fluorescens, and B. megaterium. An overnight culture of each of those bacteria was pelleted by centrifugation and dissolved in sterile Milli Q water. The cell suspensions were adjusted to an optical density of 2.0 at 620 nm by dilution in sterile Milli Q water and stored at 20°C until use (36).

DNA isolation. Total genomic DNA was purified by the 10% cetyltrimethylammonium bromide-1 M NaCl procedure used for bacterial DNA extraction (29), with the only modification that 10% N-lauroylsarcosine was used instead of 10% sodium dodecyl sulfate and a phenol extraction was included prior to the phenol-chloroform extraction.

Oligonucleotide primers and PCR amplifications. The sequences of the oligonucleotide primers used for PCR are listed in Table 2. The primers were synthesized by Eurogentec (Seraing, Belgium). The STRR and LTRR primer sequences were checked for homology to any other known sequences deposited in the available database, using the FASTA option (26). The REP and ERIC primers have been described by Versalovic et al. (35), and the PCR conditions for these primers were as specified by de Bruijn (6). For the STRR primers, the cycles were as follows: 1 cycle at 95°C for 6 min; 30 cycles of 94°C for 1 min, 56°C for 1 min, and 65°C for 5 min; 1 cycle at 65°C for 16 min; and a final step at 4°C. For the LTRR primers, the program was the same except that the annealing temperature was optimized to 45°C for 1 min with an extension at 65°C for 5 min. All the PCRs were carried out in a 25-µl volume containing 50 pmol of each primer, 1.25 mM deoxynucleoside triphosphate, 50 ng of template DNA or 1 µl of cell suspension of the control bacteria, and 1 U of DNA polymerase (DynaZyme [Oy, Espoo, Finland] for the STRR-PCR and TaKaRa [Ex Taq, Otsu, Shiga, Japan] for the LTRR-PCR). Buffers supplied with the respective enzymes were used according to the manufacturer's directions. The DNA amplification was performed in a PTC-100 Programmable Thermal Controller (MJ Research Inc., Watertown, Mass.). After the reaction, 8 µl of amplified DNA was separated on 1.5% agarose gels (Promega, Madison, Wis.), stained with ethidium bromide, and recorded with an Eagle Eye II still video system (Stratagene, La Jolla, Calif.). All PCRs were performed at least three independent times.

PCR amplification on intact filaments and cells. The cyanobacteria were pelleted by centrifugation and washed twice in sterile Milli Q water. Finally, the pellet was dissolved in an appropriate volume of sterile water to ensure that at least a few filaments or cells were present in 1 µl, which was used directly as a template for the PCR as described above. All PCRs were performed at least three independent times.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Amplification of cyanobacterial genomic DNA with PCR primers derived from repetitive sequences. Total genomic DNA, extracted from three heterocystous cyanobacteria, Nostoc strains PCC 9229 and PCC 73102 (symbiotic isolates) and Fischerella strain PCC 7521 (free-living isolate), were used as templates. The optimal PCR conditions were found at a primer-template annealing temperature of 56°C for the STRR primers and 45°C for primers LTRR 1 and 2. Furthermore, a DNA polymerase with an extended long reading capacity (see Materials and Methods) was used in the PCR with the LTRR primers. In addition, two independently processed DNA preparations from Nostoc strain PCC 9229 and Fischerella strain PCC 7521 gave the same results for each strain (data not shown). Bacterial species included as controls were R. leguminosarum biovar trifolii, P. fluorescens, B. megaterium, and E. coli, representing different phylogenetic bacterial groups. B. megaterium is a gram-positive bacterium, while the three others are gram negative, as are cyanobacteria. The use of the primer STRR 1A in the PCR on the three cyanobacteria yielded multiple distinct DNA products ranging in size from approximately 3,000 to 125 bp (Fig. 1A). Only minor PCR products were obtained from the four bacterial species included as controls. However, when the inverted primer STRR 1B was used, few PCR products were generated from the three cyanobacteria (Fig. 1B), whereas multiple distinct bands were obtained from E. coli, R. leguminosarum biovar trifolii, and P. fluorescens (Fig. 1B). The amplified PCR products obtained by using the LTRR primers ranged in size from approximately 5,000 to 500 bp (Fig. 1C). No fingerprints were obtained from E. coli, R. leguminosarum biovar trifolii, or B. megaterium, whereas one product was generated from P. fluorescens (Fig. 1C). When a combination of primer STRR 1A or STRR 1B with primer LTRR 1 or LTRR 2 at an annealing temperature of 52°C was used, only the PCR products corresponding to amplification with STRR alone were obtained (data not shown).

View larger version (41K): [in this window] [in a new window]   FIG. 1.   PCR fingerprint patterns of different cyanobacteria and other bacteria with three different primers derived from repetitive elements in cyanobacteria: Nostoc strains PCC 9229 and 73102 (symbiotic isolates; lane 1 and 2, respectively), Fischerella strain PCC 7521 (free-living; lane 3), E. coli (lane 4), R. leguminosarum biovar trifolii (lane 5), P. fluorescens (lane 6), and B. megaterium (lane 7). (A) Pattern of genomic DNA from the cyanobacteria and whole cells of the four other bacteria obtained by using primer STRR 1A; (B) pattern of the PCR product obtained by using primer STRR 1B; (C) pattern of the PCR product obtained by using the LTRR primers. Lanes C represent the control with no template DNA; lanes M are DNA molecular weight standards.

Genetic diversity among cyanobacteria. Genomic DNA was isolated from symbiotic as well as free-living cyanobacteria (Table 1). The symbiotic isolates belonging to the genera Nostoc were isolated from different host plants: Gunnera, cycads, Anthoceros, Azolla, and a lichen. The free-living cyanobacteria were represented by the genera Nostoc, Fischerella, and Synechocystis. Figure 2 shows the fingerprint patterns generated by the PCR using STRR 1A as the primer. Among the symbiotic cyanobacteria, a high diversity was found between isolates from different hosts as well as among isolates from the same hosts. However, among the 23 cyanobacteria isolated from different Gunnera species, an obvious clustering among the isolates was observed (Fig. 2A). Nostoc strain PCC 9229 and isolates 8923, 8924, 8928, and 8929, isolated from three different Gunnera species (G. monoika, G. hamiltonii, and G. chilensis [Table 1]), revealed the same fingerprint pattern (group A). The results show that a closely related group of Nostoc isolates is capable of forming symbiosis with different Gunnera species. This could further be confirmed with Nostoc isolates 8930, 8938, and 8972, isolated from G. cordifolia, G. dentata, and G. monoika, respectively, which also gave similar PCR fingerprints (group D). Nostoc isolates 8001, 8002, and 8005 (group B), isolated from G. monoika (individual plants), have identical PCR fingerprints, as do isolates 843, 891, 892, and 894 (group C) isolated from G. magellanica (individual plants). Among the four different groups, common bands (PCR products of analogous mobility) were seen with the following approximate sizes: 2,600 bp in groups B and C; 2,400 bp in groups A and B; 1,700 bp in groups A, B, and C; and 1,000 bp in groups A, C, and D (indicated by arrows in Fig. 2). In addition a fragment larger than 2600 bp was observed in groups B and C. All of the isolates belonging to groups A to D originate from New Zealand.

View larger version (47K): [in this window] [in a new window]   FIG. 2.   STRR 1A-PCR fingerprint patterns of different cyanobacteria (Table 1) with genomic DNA as the template. (A) Nostoc strains isolated from different Gunnera species. Superscript numbers refer to the Gunnera species from which they were isolated (Table 1). (B) Fingerprint patterns from symbiotic Nostoc isolated from cycads, Anthoceros sp., Azolla sp., Peltigera canina, and free-living Nostoc, Fischerella, and Synechosystis sp. (Table 1). Lanes M are DNA molecular weight standards.

View larger version (53K): [in this window] [in a new window]   FIG. 3.   PCR fingerprint patterns generated from genomic DNA of the four groups (A to D) of Nostoc isolates (Fig. 2A). (A) PCR pattern obtained with STRR 1B as primer; (B) PCR pattern obtained with LTRR 1 and 2 as primers. Lanes M are DNA molecular weight standards.

View larger version (57K): [in this window] [in a new window]   FIG. 4.   PCR fingerprint patterns of nonheterocystous cyanobacteria obtained with primer STRR 1A. Intact filaments and cells were used as material. Lane C represents the control with no template DNA; lanes M are DNA molecular weight standards.

DNA fingerprinting of intact cyanobacterial filaments. To simplify the procedure and to make it more useful on natural populations, intact cyanobacterial filaments or cells were used directly in the PCR. A number of free-living and symbiotic isolates were tested. Paired comparisons of the PCR amplification products from purified DNA and from filaments are shown in Fig. 5. The same fingerprint pattern was obtained whether purified DNA or intact filaments were used. Furthermore, the age of the culture (1 to 3 weeks) did not influence the fingerprint pattern (data not shown).

View larger version (57K): [in this window] [in a new window]   FIG. 5.   Comparison of fingerprint patterns of intact filaments with extracted genomic DNA, using primer STRR 1A. Lanes with odd numbers represent the reaction where genomic DNA was used as the template, and even-numbered lanes represent the reactions on filaments. Lanes M are DNA molecular weight standards.

ERIC- and REP-PCR fingerprint of chromosomal DNA from axenic cyanobacterial isolates. The use of primers derived from the common repetitive sequences found in most gram-negative bacteria, ERIC and REP, were investigated on chromosomal DNA extracted from axenic cyanobacterial isolates (Nostoc isolates 8001, 8002, and 8005 and strains identified with a PCC number in Table 1). The results show that both ERIC and REP sequences generated distinct PCR profiles in the cyanobacteria investigated in this study (Fig. 6). A high diversity among the cyanobacteria tested except for Nostoc strains 8001, 8002 and 8005 was observed with both primers. By the use of ERIC primers, fingerprints of Nostoc isolates 8002 and 8005 were similar whereas that of Nostoc isolate 8001 differed (Fig. 6A). However, only minor differences in fingerprint profile among the three isolates were obtained with the REP primers (Fig. 6B). The axenic cyanobacteria showed different fingerprint patterns with all the three primers, STRR, ERIC, and REP (Fig. 2 and 6).

View larger version (71K): [in this window] [in a new window]   FIG. 6.   PCR fingerprint patterns of different cyanobacteria (symbiotic and free-living) based on extracted DNA from axenic cultures (Table 1). (A) PCR pattern generated with ERIC primers; (B) pattern generated with REP primers. R. leguminosarum biovar trifolii was included as a positive control. Lanes M are DNA molecular weight standards.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In this study, we have demonstrated that a PCR fingerprinting method based on the presence of STRR and LTRR sequences present in the cyanobacterial genome can be used as a genetic tool for identification and diversity studies of cyanobacteria. The method was shown to be accurate in distinguishing and classifying even closely related strains. The described method is a valuable and rapid alternative to other methods used for classification and diversity studies of cyanobacteria. In contrast to RFLP analysis (20, 32, 37), DNA extraction, Southern blotting, and probe hybridization are not required. Moreover, PCR fingerprint patterns can, as demonstrated here, be obtained directly from intact filaments and cells.

The distribution and occurrence of STRR sequences among cyanobacteria have been investigated by Southern blot analysis (23). Hybridization was found in all heterocystous cyanobacteria that were tested except the free-living Nostoc strain PCC 6720, whereas filamentous nonheterocystous and unicellular cyanobacteria, including Synechocystis, did not show hybridization. Among all of the heterocystous cyanobacteria used in this study, Nostoc strain PCC 6720 was the only one that did not generate a PCR fingerprint, confirming the lack of STRR sequences in this strain as observed by Mazel et al. (23). A distinct PCR fingerprint was observed in nonheterocystous cyanobacteria, including Synechocystis (23). Based on this observation, conclusions on the presence of STRR sequences as tandemly repeated elements in nonheterocystous cyanobacteria cannot be drawn. The fingerprint pattern is only an indication that the sequence used as primer is present in the genome. The STRR elements were found in some strains of the nonheterocystous cyanobacterium Microcystis, although it was concluded that few copies are present in the genome (32). In the same study, the STRR element was successfully used as a probe to characterize and classify free-living Nostoc and Anabaena strains (32). The STRR-PCR fingerprint method was used on 30 symbiotic Nostoc isolates, with the majority of isolates originating from different Gunnera species (1). The results revealed both a high genetic diversity among the isolates and a distinct clustering (Fig. 2A). Based on the technique used in this study, the individual isolates in each group must be considered as similar or closely related. A similar fingerprint pattern was obtained from the axenic isolate Nostoc strain PCC 9229 and four nonaxenic isolates collected from three different Gunnera species (group A). The identical fingerprint patterns obtained from those isolates indicate that the developed method can be used on nonaxenic isolates, collected directly from the symbiotic tissue. Moreover, the result supports earlier observations that one strain or closely related strains can form symbiosis with different Gunnera species (36) and, based on reconstitution experiments, even with different plant groups (4, 7, 13). Nostoc isolates 8001, 8002, and 8005 (group B), isolated from individual plants of G. monoika, have previously been used to examine diversity by RFLP analysis, and it was concluded that Nostoc isolates 8002 and 8005 were most likely identical (37). In this study, the three isolates have identical fingerprint patterns, using the STRR primers. However, the results with LTRR and ERIC primers show identical patterns only with Nostoc isolates 8002 and 8005, as demonstrated by Zimmerman and Bergman (37). This finding indicates that the various primer sets have different degrees of resolution, and in order to draw a more specific conclusion about diversity or similarity among closely related isolates, different primers have to be included in the PCR analysis. This observation was also evident by comparing the clustering of the isolates into four groups by the STRR and LTRR primers. STRR 1A and STRR 1B revealed the same clustering. The limited number of PCR products obtained with STRR 1B primer in contrast to STRR 1A might be a reflection of the position and orientation of the individual STRR sequences in the cyanobacterial genome. However, with the LTRR primers, some differences were obtained among individuals within a group. The difference in banding patterns of isolates 8923, 8928, and 892 compared to the other isolates in their respective groups was due to an absence of PCR product in the high-molecular-weight range.

The use of rep-PCR for fingerprinting and diversity studies has been shown to be a powerful technique for many bacteria, i.e., Rhizobium species and other soil bacteria (6, 17, 25), Xanthomonas and Pseudomonas species (18), Bartonella species (31), and Legionella species (9). All of these studies are based on the distribution of the widespread REP and ERIC sequences among eubacteria, primarily in the gram-negative group. In a previous study where various bacteria were screened for the presence of these sequences, it was shown that only the ERIC sequence was present in cyanobacteria, represented by Anabaena sp. (35). In our study, we have demonstrated that both ERIC- and REP-derived oligonucleotides produce fingerprints of cyanobacteria. Both sequences gave distinct reproducible PCR profiles and can therefore be used as primers for genomic fingerprinting of cyanobacteria. However, due to the common presence of these sequences in many bacteria, the use of ERIC and REP primers require axenic cultures. As it can be difficult and time-consuming to establish axenic cultures of cyanobacteria, the developed PCR method with STRR or LTRR primers provides a useful method for studying the diversity of cyanobacteria in the natural environment, whether free-living or symbiotic. Moreover, an important and useful result obtained in this study is the application of the fingerprint method directly on intact cyanobacterial filaments and cells. Although most of the symbiotic isolates are surrounded by a polysaccharide sheath, the PCR profiles were indistinguishable from those generated with purified DNA. The use of intact cells for PCR represent a very efficient way of analyzing bacterial isolates and a method which has been applied on different bacterial species (14, 36).