OUP user menu

Evidence for polyphyletic origin of the members of the orders of Oscillatoriales and Pleurocapsales as determined by 16S rDNA analysis

Tatsuya Ishida , Makoto M. Watanabe , Junta Sugiyama , Akira Yokota
DOI: http://dx.doi.org/10.1111/j.1574-6968.2001.tb10736.x 79-82 First published online: 1 July 2001


Phylogenetic analyses were carried out on six pleurocapsalean strains and 12 oscillatorialean strains by sequence determination of the 16S rDNA (16S rRNA gene). Although heterocyst-forming strains of the orders Nostocales and Stigonematales were shown to be monophyletic, unicellular strains of the orders Chroococcales and filamentous Oscillatoriales were shown to be polyphyletic, as reported earlier. Moreover, unicellular and baeocyte-forming strains of the order Pleurocapsales, which were thought to be monophyletic, were newly found to be polyphyletic. The results strongly indicate that even morphology was not necessarily reflected in the phylogenetic relationships at the order level. A need exists to reconstruct the taxonomy of cyanobacteria at the order level.

  • Cyanobacterium
  • Pleurocapsales
  • 16S rDNA
  • Phylogeny

1 Introduction

Cyanobacteria are autotrophic bacteria that perform oxygenic photosynthesis. The taxonomy of cyanobacteria has until now been based mainly on their morphology, and cyanobacteria are now separated into five orders [1]: Chroococcales (I), Pleurocapsales (II), Oscillatoriales (III), Nostocales (IV), and Stigonematales (V). Phylogenetic analyses of cyanobacteria based on 16S rDNA were carried out by several groups [26], which indicated that Chroococcales (I) and Oscillatoriales (III) were polyphyletic. Other orders, Nostocales (IV) and Stigonematales (V), which form heterocysts, were shown to be monophyletic. Of the other orders, monophylicity was indicated for Pleurocapsales (II) which forms internal spores (baeocytes); however, only a few strains were phylogenetically analyzed. Furthermore, the sequences determined in pleurocapsalean strains (II) were incomplete. As in the other orders of cyanobacteria, the taxonomy of Pleurocapsales (II) was based on its morphology [7]. The morphology of cyanobacteria is known to change easily and is not conserved [8].

To grasp a general image of Pleurocapsales (II) and to get more information on Oscillatoriales (III), in this study we investigated the phylogenetic position of the six pleurocapsalean strains (II) and 12 oscillatorialean strains (III) based on 16S rDNA sequences, which have been proved to be very useful in the study of bacterial phylogeny [9].

2 Materials and methods

2.1 Cyanobacterial strains

The strains examined were as follows: Chroococcidiopsis sp. PCC 6712 and C. thermalis PCC 7203; Stanieria sp. PCC 7301 and S. cyanosphaera PCC 7437T; Pleurocapsa sp. PCC 7319 and PCC 7327; Geitlerinema sp. PCC 7105; Leptolyngbya sp. PCC 7104 and PCC 7375; Lyngbya aestuarii PCC 7419T; Oscillatoria acuminata PCC 6304 and O. sancta PCC 7515T; Pseudanabaena sp. PCC 6802, PCC 6903, PCC 7367, PCC 7403, and PCC 7408; Symploca sp. PCC 8002. All strains were obtained from the Pasteur Culture Collection (PCC), and all were axenic. Pleurocapsa sp. PCC 7319 and Pseudanabaena sp. PCC 7367 were grown in medium ASN-III [10], Geitlerinema sp. PCC 7105, Leptolyngbya sp. PCC 7375, and Stanieria sp. PCC 7301 in medium ASN-III.B12 [10], Symploca sp. PCC 8002 in medium ASO-III [10], Chroococcidiopsis sp. PCC 6712, C. thermalis PCC 7203, Leptolyngbya sp. PCC 7104, O. acuminata PCC 6304, O. sancta PCC 7515T, Pleurocapsa sp. PCC 7327, Pseudanabaena sp. PCC 6802, PCC 6903, PCC 7403, PCC 7408, and S. cyanosphaera PCC 7437T in medium BG-11 [10], and L. aestuarii PCC 7419T in medium MN [10]. These strains were cultured by incubation in test tubes without aeration at 20°C with fluorescent lamp illumination of 500 lux. The names of strains used fundamentally followed the catalog of PCC [10].

2.2 PCR amplification and sequencing

The methods for DNA extraction, amplification, and sequencing of 16S rDNA were the same as those described previously [3], except for PCR oligonucleotide primers and conditions. The almost complete 16S rDNAs from the genomic DNA of the respective strains were amplified by PCR, using oligonucleotide primers by the combination of 1 (5′-AGAGTTTGATCCTGGCTCAG-3′) and 16 (5′-ACGCCGACCTAGTGGAGGAA-3′) [11]. PCR was performed at first for 3 min at 94°C and then 30 cycles with the following features: 1 min at 94°C, 1 min at 55°C, and 2 min at 72°C, followed by a final elongation step for 10 min at 72°C.

2.3 Alignment and phylogenetic analyses

The new 16S rDNA sequences were multiple-aligned using CLUSTAL W, version 1.74 [12] with a selection of cyanobacterial reference sequences obtained from the DNA Data Bank of Japan (DDBJ). The alignment was corrected manually and converted to a distance matrix. A phylogenetic tree was constructed from the distance matrix data by applying the algorithm of the neighbor-joining (NJ) method [13] to Knuc values [14] with multiple substitutions corrected and positions with gaps excluded. To evaluate the robustness of branches in the inferred tree, the bootstrap resampling method of Felsenstein [15] with 1000 replicates was used. Maximum likelihood (ML) analysis was also carried out using the MOLPHY program, version 2.3b3 [16] based on the same alignment data after positions with gaps were excluded. The ML distance matrix was calculated using NucML, and the initial NJ tree was reconstructed by NJdist in the MOLPHY package. The ML tree was finally obtained using NucML with the local rearrangement method from the NJ tree with the HKY model, and local bootstrap probabilities were estimated by a resampling of the estimated log-likelihood method [17].

3 Results and discussion

The new 16S rDNA sequences were deposited in DDBJ; accession numbers for each strain are given in Fig. 1. Sequences from 58 cyanobacterial strains and two Gram-positive bacterial strains (Bacillus subtilis subsp. subtilis DSM 10 and Clostridium butyricum ATCC 19398) as the outgroups were aligned. Positions with gaps and undetermined and ambiguous sequences were removed. A total of 1214 sites were used for the phylogenetic analysis.

Figure 1

ML tree of 58 cyanobacterial strains and two Gram-positive bacterial strains (Bacillus subtilis subsp. subtilis DSM 10 and Clostridium butyricum ATCC 19398) as the outgroups. An alignment of 1214 nucleotides after excluding the positions with gaps was used. Scale bar=1 base substitution per 10 nucleotide positions. Local bootstrap probabilities are indicated at nodes if larger than 80. The strains of which the 16S rRNA genes were determined in this study are indicated in bold. Accession numbers in the DDBJ, EMBL, and GenBank databases are in parentheses. Order VI: Prochlorales; T: type strain.

Constructed phylogenetic trees, the ML tree (Fig. 1) and the NJ tree (not shown), revealed that all the nostocalean and stigonematalean strains, which are differentiated by the formation of heterocysts in their trichomes, were monophyletic. The unicellular strains, Chroococcales (I) and Pleurocapsales (II), were indicated, as the filamentous oscillatorialean strains (III) were, to be polyphyletic and did not form heterocysts. In previous reports [25], except for the pleurocapsalean strains (II), the same results were obtained, and they indicated that the pleurocapsalean strains (II) were monophyletic; however, only a few pleurocapsalean strains (II) were analyzed in that report. In this study, we included more pleurocapsalean strains (II) in the phylogenetic analyses. As shown in Fig. 1, these strains are clearly separated into at least three clusters. The major one consists of six pleurocapsalean strains (II) and one prochloralean strain (VI). The others consist of only one pleurocapsalean strain (II) and strains from other orders; one was an oscillatorialean strain (III) and the other was a chroococcalean strain (I). In this study, some new sequences were added to cyanobacterial phylogenetic analysis, and it became clear that the order Pleurocapsales (II) is polyphyletic. That is, pleurocapsalean strains (II) of which the phylogenetic positions have been analyzed until now were thought to be placed in a narrow range. Namely, only a part of the cyanobacteria were examined in the phylogenetic analyses, and these analyses were thought to be insufficient under the present conditions. At present, there are only Nostocales (IV) and Stigonematales (V) of which the morphological characteristics as the guideline to the order were consistent with the phylogeny (Fig. 1). Cyanobacteria seemed to have diverged at the same time in an early period in the cyanobacterial phylogenetic tree [26], and it is difficult to elucidate the process of their evolution. This indicates that diversification of cyanobacteria once occurred during a certain early period. As a result, chroococcalean strains (I) and oscillatorialean strains (III) might be seen in the same cluster on some phylogenetic position in the present phylogenetic trees. It is interesting, however, that the chroococcalean strain (I) Synechococcus sp. PCC 7002 and the oscillatorialean strain (III) Oscillatoria rosea M-220 probably diverged recently. Thus unicellular and filamentous strains are close, and it may help in the determination of morphological change from unicellular to filamentous strain or filamentous to unicellular strain.

In this study, we examined not only pleurocapsalean strains (II), but also oscillatorialean strains (III). The 16S rDNA sequences of 12 oscillatorialean strains (III) were determined and their phylogenetic positions analyzed. As shown in Fig. 1, oscillatorialean strains (III) are separated into eight clusters supported by high statistical reliability. Three groups consisted of a few oscillatorialean strains (III), and therefore the sequence of 16S rDNA of more strains should be determined. Although almost oscillatorialean strains (III) did not form a consistent cluster even at the genus level, only the Pseudanabaena species characterized by the presence of a cross-wall [18] formed a single cluster. Thus, all the morphological characteristics at the genus level were not always significant. Namely, a taxonomic system of cyanobacteria reflecting phylogenetic relationships should be reconstructed.


This work was supported by Research Fellowships of the Japan Society for the Promotion of Science for Young Scientists.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
View Abstract