Correspondence: Otakar Strunecký, Institute of Botany, Academy of Science of the Czech Republic, Dukelská 135, Třeboň, Czech Republic. Tel.:+420 384721156; fax: +420 384 721 136; e-mail: firstname.lastname@example.org
Cyanobacteria are well adapted to freezing and desiccation; they have been proposed as possible survivors of comprehensive Antarctic glaciations. Filamentous types from the order Oscillatoriales, especially the species Phormidium autumnale Kützing ex Gomont 1892, have widely diverse morphotypes that dominate in Antarctic aquatic microbial mats, seepages, and wet soils. Currently little is known about the dispersion of cyanobacteria in Antarctica and of their population history. We tested the hypothesis that cyanobacteria survived Antarctic glaciations directly on site after the Gondwana breakup by using the relaxed and strict molecular clock in the analysis of the 16S rRNA gene. We estimated that the biogeographic history of Antarctic cyanobacteria belonging to P. autumnale lineages has ancient origins. The oldest go further back in time than the breakup of Gondwana and originated somewhere on the supercontinent between 442 and 297 Ma. Enhanced speciation rate was found around the time of the opening of the Drake Passage (c. 31–45 Ma) with beginning of glaciations (c. 43 Ma). Our results, based primarily on the strains collected in maritime Antarctica, mostly around James Ross Island, support the hypothesis that long-term survival took place in glacial refuges. The high morphological diversification of P. autumnale suggested the coevolution of lineages and formation of complex associations with different morphologies, resulting in a specific endemic Antarctic cyanobacterial flora.
Older Antarctic studies focusing on the origin of terrestrial life in Antarctica, with its extensive glaciations during the past such as the one summarized by Convey et al., 2008; implicated the impossibility of survival of organisms on land, especially during the glacial maxima (Huybrechts, 1993; Denton & Hughes, 2002). However, new and more precise knowledge of glaciations, together with the rising amount of molecular data of various kinds of organisms, is changing our understanding of the ability of living organisms to persist during extensive glaciations of Antarctica.
Cyanobacteria are distributed worldwide in both aquatic and terrestrial environments. They inhabit extreme environments including various polar habitats (Whitton & Potts, 2000). Cyanobacteria are well adapted to freezing and desiccation (Šabacká & Elster, 2006), and they have been proposed as possible survivors of comprehensive Antarctic glaciations (Vincent et al., 2004). Filamentous types from the order Oscillatoriales, especially the species Phormidium autumnale Kützing ex Gomont 1892, sensu lato have widely diverse morphotypes that dominate in Antarctic aquatic microbial mats, seepages, and wet soil (Broady, 1996; Komárek et al., 2008). Various specimens of P. autumnale sensu stricto were found inhabiting different habitats at James Ross Island, Antarctica (Komárek et al., 2008; Strunecký et al., 2010). There is a long-lasting discussion about the dispersal of cyanobacteria to Antarctica (Mullins et al., 1995; Wilmotte et al., 1997; Komárek, 1999; Nadeau et al., 2001; Convey et al., 2008). Reintroduction or persistence of cyanobacterial species during cold climatic cycles connected with the extensive ice cover cleaning action may be confirmed by molecular studies of collected strains within Antarctica and by their relationship with populations on surrounding islands and continents.
Evidence of endemic species in Antarctica based on ‘molecular clocks’ can be seen through several traits of diverse kinds of various organisms. Allegrucci et al. (2006) showed that chironomids species from the Antarctic Peninsula separated from its close relative on South Shetland Island before 49 Ma, while both separated from their south American relatives 69 Ma. Springtails and mites are reported to have survived glacial maxima of Pleistocene and Holocene in refugee of Dry Valleys (McGaughran et al., 2008). Maslen & Convey (2006) hypothesized that nematode fauna of the Alexander Island in the western part of Antarctic Peninsula survived at least through the Pleistocene glaciation. Analogous results are reported for eukaryotic green microalgae (De Wever et al., 2009), where estimated divergence times of Antarctic species were between 17 and 84 Ma and the lineages with long branches evolved between 330–708 Ma. Another occurrence of Antarctic endemism in eukaryotic microalgae was introduced by Rybalka et al. (2009), who examined the differences in genetic diversity of Tribonemataceae (Xanthophyceae) from Antarctic and from European (temperate) areas. In psbA/rbclL spacer region, the differences between Antarctic and temperate strains were greater than those within either Antarctic or temperate strains. Similar results were obtained by Bahl et al. (2011) on Chroococcidiopsis (Chroococcales, Cyanobacteria) by numerous pyrosequencing of environmental samples, which confirmed that Chroococcidiopsis variants were specific to either hot or cold deserts. These findings imply that the Antarctic and temperate strains represent at least two different populations of a single species. Studies based on morphology of diatoms suggest that in some areas at least 40% of species were endemic to Antarctica (Sabbe et al., 2003; Vanormelingen et al., 2008; Kopalová et al., 2009). These high levels of endemism suggest the importance of relatively low dispersal rates and long-term survival in isolated refuges. However, a large portion of this endemism validation, except for Chroococcidiopsis (Bahl et al., 2011), concerns eukaryotic organisms, and the molecular evidence for Antarctic endemism of prokaryotes (cyanobacteria) is still under discussion.
Antarctica was located at the current position during the mid-Cretaceous (Lawver & Gahagan, 2003). The breakup of Antarctica and South America occurred around 31–45 Ma (Million Years Ago) (Lawver & Gahagan, 2003; Eyles, 2008). This event was tightly linked to rapid Cenozoic glaciations at the Eocene/Oligocene boundary (c. 34 Ma) (DeConto & Pollard, 2003). At that time, the origin of the Drake and Tasmania passages facilitated the creation of the Antarctic Circumpolar Current and Polar Frontal Zone that are considered to be the main barriers to the dispersal of organisms to Antarctica. In the middle Miocene (c. 15 Ma), gradual cooling caused the reestablishment of permanent East and West Antarctic ice sheets, which persisted continuously until c. 10 Ma (Zachos et al., 2001).
We present the first study of Antarctic prokaryotic photosynthetic microorganisms from the order Oscillatoriales, specifically the species P. autumnale, which proposed that a long-term survival in Antarctica is possible. We used two different approaches of divergence time estimations for cyanobacteria: (1) auto-correlated rate of evolution (Sanderson, 2003) and (2) the Bayesian evolutionary analysis with strict molecular clocks (Drummond & Rambaut, 2007). The divergence times for the estimation were based on Tomitani et al. (2006) analysis of geochemical, geological, and phylogenetic data that provided an internal bacterial calibration point, that is, the cyanobacteria that were able to form heterocytes diverged between 2450 and 2100 Ma. This allowed us to calculate divergence times of 12 strains of P. autumnale from James Ross Island and from six other sites within Antarctica.
Materials and methods
Phormidium autumnale specimens were collected in the vicinity of Antarctic Peninsula, at James Ross, King George, and Killingbeck Islands, Antarctica (Table 1). The isolation and cultivation of strains were described elsewhere (e.g. Elster et al., 1999; Strunecký et al., 2010). The isolated strains were identified according to Komárek & Anagnostidis (2005). Strain morphologies were analyzed using optical microscope (Olympus BX 51) (Fig. 1), and the widths and lengths of at least 50 cells were measured for each Phormidium strain under 1000× magnification. Statistical analysis was made by statistica 9 (Statsoft); t-test for independent samples was used for comparison of measured parameters.
Table 1. Origin of the cyanobacterial strains examined in this study belonging to species Phormidium autumnale sensu stricto
The DNA extraction protocol was based on Smalla et al. (1993) after the method of Taton et al. (2003). A 0.5 mg of sample was suspended in 0.5 mL of SNT solution (500 mM Tris–HCl, 100 mM NaCl, 25% saccharose) supplemented by 26 μL of freshly added lysozyme (25%, w/v). The resulting suspension was shaken and incubated for 30 min at 37 °C. After incubation, 0.5 mL of solution II (500 mM Tris base, 500 mM EDTA, 1% sodium dodecyl sulfate, 6% phenol) and 0.25 g of glass beads (diameter, 0.17–0.18 mm; Braun Biotech) were added to the sample and shaken for 2 min on shaker. The resulting suspension was placed on ice for 1 h and vortexed every 10 min. After incubation, the suspension was centrifuged for 10 min at 720 g (Eppendorf), and 1 mL of the aqueous phase was mixed with an equal volume of phenol, after which it was centrifuged for 5 min at 13 600 g. The supernatant was then transferred into new tubes, extracted with equal volumes of phenol–chloroform–isoamyl alcohol (25 : 24 : 1), and reextracted with equal volumes of chloroform–isoamyl alcohol (24 : 1). Subsequently a standard Na acetate–ethanol precipitation was performed, and the dried pellet was resuspended in 100 μL of TE buffer (10 mM Tris–Cl, 1 mM EDTA; pH 8). Cyanobacteria of species P. autumnale contains one copy of 16S rRNA gene and its adjacent 16S–23S internal transcribed spacer (ITS) fragment (Lokmer, 2007); hence, we directly continued PCR. The 16S rRNA gene with the 16S–23S intergenetic segment was amplified using primers 359F (GGGGAATYTTCCGCAATGGG (Nübel et al., 1997)) and 23S30R (CTTCGCCTCTGTGTGCCTAGGT) (Wilmotte et al., 1993) with the following settings: starting denaturalization step (94 °C, 5 min); 40 cycles of 30 s at 94 °C, 30 s at 53 °C, and 3 min at 72 °C; final extension for 7 min at 72 °C, and cooling to 4 °C. Successful PCR was confirmed by running a subsample on a 1% agarose gel stained with ethidium bromide. PCR products were purified using a QIAquick PCR Purification Kit. Sequencing of the 16S rRNA gene fragment was performed on an ABI 3130 sequencer, using BD3.1 (Applied Biosystems) chemistry, with six primers (359F, 23S30R, CYA_1064R – GATTCGCGACATGTCAAGTCTTGGTAAGG, CYA783F-TGGGATTAGATACCCCAGTAGTC, S17*-GGCTACCTTGTTACGAC, and ILE23F – ATTAGCTCAGGTGGTTAG (Wilmotte & Herdman, 2001; Strunecký et al., 2010) to obtain complementary sequences. The monophyletic origin of studied strains was evaluated by phylogenetic comparison under extensive evaluation of more than 1300 sequences of Oscillatoriales cyanobacteria available at GenBank (http://www.ncbi.nlm.nih.gov) and at the Ribosomal database project (rdp.cme.msu.edu) (unpublished). For the examination of molecular phylogeny, 14 sequenced strains (GenBank accession numbers JN230326–JN230347) were combined with the strains from our previous study (Strunecký et al., 2010) to obtain comparisons to non-Antarctic strains. Sequences were aligned by Mafft 6 (Katoh & Toh, 2010) and inspected in Bioedit 7.0 (Hall, 1999). Phylogenetic relationship of P. autumnale based on partial 16S rRNA gene and ITS data (length c. 1681 bp starting at position 345, Escherichia coli IHE3034 16S rRNA gene) was calculated using the neighbor-joining analysis in mega5 (Tamura et al., 2011). Topology vas validated through the Bayesian analysis in MrBayes 3.1.2. (Huelsenbeck & Ronquist, 2001) at http://www.metacentrum.cz. For the Bayesian analysis, two runs of four Markov chains with over 1 000 000 generations, and sampling every 100 generations, were employed. The initial 250 000 generations were discarded as burn-in.
To estimate the timing of phylogenetic divergence events, we used a partial 16S rRNA gene fragment of length c. 1011 bp (E. coli bp 345–1356). The GenBank review of Antarctic data and blast of 16S rRNA gene sequences resulted in seven Antarctic matches of corresponding position and length, as we had used all suitable Antarctic P. autumnale sequences with known place of origin. Additional two strains from Siberia, for the comparison of polar biotopes, and New Zealand, for the comparison of geographically proximate site, were included. Phylogenetic estimation for relaxed molecular clocks was made through Dnamlk in Phylip 3.7 (Felsenstein, 2005). Relative node ages were estimated using R8S 1.7 (Sanderson, 2003) utilizing penalized likelihood method and smoothing parameter of one thousand. Strict molecular clocks were estimated using beast 1.6.1 (Drummond & Rambaut, 2007) using the HKY substitution model with invariant and four gamma categories of sites with gamma shape = 0.17 obtained by ModelTest (Posada & Crandall, 1998); tree speciation of the Yule process was used with 10 000 000 of MCMC states logged every 1000th generation. Chronotree was constructed in TreeAnnotator within beast package with posterior probability limit of 0.5 by combining three MCMC runs with previous burning of 2500 for every run. Final trees were made by Figtree 1.31 (Drummond & Rambaut, 2007) and graphically arranged in Adobe Illustrator.
Time calibration for both methods was performed under the assumption that heterocytous cyanobacteria, which are a monophyletic phylogenetic clade (Schirrmeister et al., 2011), should be formed in times of the great oxygenation event 2100–2450 Ma (Tomitani et al., 2006). The split of Chroococales (Prochlorococcus marinus MIT 9303, Synechococcus elongatus PCC 7942) and Nostocales (Anabaena flos-aquae UTCC 64, Nostoc sp. PCC 7120) at 2250 Ma, as representatives of heterocyte-deficient and heterocyte-carrying cyanobacteria, respectively, was used as two outgroups for calibration.
The microscopic observation of the morphology of the studied strains (Fig. 1) confirmed their classification as the cyanobacterial morphospecies Phormidium autumnale (Oscillatoriales, Cyanobacteria). Morphology (difference in trichome widths) evaluation was completed utilizing six strains from clusters I, II, and IV (Fig. 2). All Antarctic strains chosen as representatives of clusters, except for the JR3 strain, had statistically different widths of cells in trichomes (P < 0.05). The constraints of the width of trichomes are shown in Fig. 3.
The phylogenetic tree of strains of P. autumnale based on the 16S rRNA gene and 16S–23S rRNA ITS fragment formed four distinct clusters of strains originating at James Ross, King George, and Killingbeck Islands, Antarctica (Fig. 3). A sequence identity matrix (data not shown) showed that there was a 9% difference in 16S rRNA gene combined with 16S–23S ITS rRNA gene among the most diverse Antarctic strains JR5 and JR3, whereas the largest difference of the 16S rRNA gene was 3% between the strains KI19 and JR3; the biggest difference of 16S–23S ITS rRNA gene of 18% was between strains JR5 and two other strains JR3 and JR6.
Estimation of divergence times in R8S for 16S rRNA gene fragment (Fig. 4a) revealed that speciation of P. autumnale strains started in 435 Ma, the following diversification events were found in 400 Ma, and more intense speciation was found between 311 and 284 Ma. Two speciation events can be found in 232 Ma, three between 166 and 144 Ma, and another two in 64–62 Ma period. Between 35 and 32 Ma, three events can be found. The most recent change was occurred 0.32 Ma. The date of diversification of P. autumnale of Antarctic strains from those outside of this continent was calculated to be at 90 Ma for the Siberian strain, 155 Ma for the European and Japan strains, 166 Ma for the New Zealand strain, and 284 and 303 Ma for the strains from North America (Canadian Arctic) and Europe, respectively.
beast strict molecular clocks estimation (Fig. 4b) revealed that speciation of P. autumnale strains started 297 Ma, the following splits were found at c. 236 to c. 167 Ma with separation of the Arctic strain between 128 and 119 Ma. Diversification of the European and North American strains was found between 94 and 66 Ma with common speciation events of the Antarctic strains between 65 and 61 Ma. The Siberian strain using beast estimation diverged in 54 Ma. A hot spot of four events in speciation of the Antarctic strains was found between 38 and 31 Ma, with first speciation of the Forlidas strains from the Transantarctic Mts. at the beginning of this period.
The Antarctic microflora, particularly the cyanobacterial types, is often considered as belonging to cosmopolitan cold-adapted ecotypes, which have been commonly transported to Antarctica from other continents by various means of transport, most likely by winds and birds (Marshall, 1996; Marshall & Chalmers, 1997; Jungblut et al., 2010; Namsaraev et al., 2010). The idea of occurrence of various endemic species of the Antarctic was already discussed by several authors (Sabbe et al., 2003; De Wever et al., 2009; Vyverman et al., 2010), where endemism was postulated on the basis of variability in DNA or because of different morphology of algae. The morphological differences of various Antarctic cyanobacteria with their counterparts found outside of Antarctica are obvious (Komárek, 1999; Komárek et al., 2008) Additionally, previous molecular study of Antarctic cyanobacteria using 16S rRNA gene suggested that isolates had a long association with the Antarctic location (Taton et al., 2003, 2006; Casamatta et al., 2005) More recent molecular analyses of Antarctic cyanobacterial populations with greater possibilities of comparison with the ever growing gene database have not found even one totally identical genotype in the 16S rRNA gene composition from other continents (Taton et al., 2006, 2011; Jungblut et al., 2010; Strunecký et al., 2010, 2011).
We used a different approach to determine the endemism of Antarctic cyanobacteria based on evolution of studied species on that time scale. Clear fossil morphological specimens with well-defined assignation with taxonomical units of current cyanobacteria do not exist. The absence of such fossil cyanobacteria needed different methodology. Therefore, we used as calibration point for our models the great oxygenation event between 2450 and 2100 Ma (Tomitani et al., 2006). The arbitrary calibration point for analyses, 2250 Ma, in the middle of this period was set deep in the past, and hence, the accuracy of modeled divergence times should not be absolute. The exact divergence times could be placed somewhere between these two most extreme modeled limits that emerged from extensive calculation with various initial model settings. The start of evolution within P. autumnale was found within our R8S chronotree deep in the Paleozoic, 442 Ma; in beast, chronotree was found later, 297 Ma. We calculated that European and North American strains diverged from the Antarctic ones between 197 and 145 Ma with R8S and 166–85 Ma according to beast, respectively (Fig. 4). The isolation of Antarctic strains continues from the time of the breakup of Antarctica and South America in c. 30–45 Ma. Four speciation events in beast between 38 and 30 Ma and two in R8S around 32 Ma, respectively, were found for the studied strains. Such results are confirmed by the final split of Antarctica from the south of Tasmania and south of South America with the rise of glaciations in Antarctica before c. 33–34 Ma (DeConto & Pollard, 2003; Barker et al., 2007) and sustained persistence of ice sheets (Zachos et al., 2001). Those speciation events indicated rapid evolution of lineages capable of coping with the more severe climate. Divergence times for strains from the Forlidas Valley, which appeared 3 and 3–8 Ma before divergence events of James Ross Island strains in 32 and 38 Ma in R8S and beast models, were in good correspondence with earlier glaciation of the Transantarctic Mountains, which were partially glaciated until 43 Ma (Eyles, 2008).
The analysis of Liu et al. (2009) indicated that some nunataks could be partially ice free, even if their tops were covered by ice during the Pleistocene ice ages. The nunataks of the continental Antarctica remained ice free at least for the last glacial maximum (Hall, 2009). The nunatak glacier boundary is the suitable place for the growth of the P. autumnale, as was evidenced by the authors in continental Antarctica, which is also a suitable place for the survival of other types of cyanobacteria (Namsaraev et al., 2010). Geological data for the last 6.2 Ma indicate that James Ross Island was covered by relatively thin, wet-based glaciers that overlaid the tops of the mesas by at the most 45–200-m-thick layer of ice (Smellie et al., 2008). Other refuges such as frozen lakes where cyanobacteria could possibly survive (Singh & Elster, 2007) can be found. Also, tidal zones of maritime land (Furnes & Fridleifsson, 1978) or seepages on the seashore possibly remained ice free and could have served as suitable areas for cyanobacterial life. Because the genotypes of P. autumnale from continental Antarctica and other parts of the Antarctic Peninsula differed from the P. autumnale found at James Ross Island, the most straightforward explanation was that cyanobacterial population of this genera survived somewhere in locally suitable refuge.
Gould & Eldredge (1993) suggested a process of very slow (static) genetic change with genetic drift in small, sexually reproductive populations or concerted evolution with exchange of genetic material for neutrally evolving genes and/or neutral mutations. This theory was tailored for cyanobacteria where it was called the hypobradytelic evolution – that is, morphology and physiology of cyanobacteria have changed little or not at all over thousands of millions of years (Schopf & Packer, 1987; Siefert & Fox, 1998; Kremer, 2006; Schirrmeister et al., 2011). The genetic exchange in asexually dividing identical cells of mats and connected cells in trichomes of cyanobacterial populations should, in principle, contradict genetic drift. Fossil remains of tubular microfossils that can be attributed to modern Oscillatoriales were found as far as the Proterozoic (Schopf & Walter, 1982). For example, trichomes of microfossils from Gaoyushuang formation in Northern China 1425 Ma (Schopf, 1988) and from Lakhanda formation in Eastern Siberia 950 Ma (Whitton & Potts, 2000) are morphologically indistinguishable from modern forms of Phormidium. We are not disputing that these old microfossils are relatives of modern Ocillatoriales, although our data confirm the evidence of their low rate of evolution. The slow rate of evolution of cyanobacteria in general shall be confirmed by further studies. However, our findings in Antarctica suggest that Phormidium-like cyanobacteria were convenient for morphological and physiological adaptation to physical pressure of the environment without significant changes in ribosomal DNA. The ‘hypobradytelic’ theory strengthens the hypothesis that roots of cyanobacteria belonging to Oscillatoriales could be quite old, and the biogeographic history of Antarctic cyanobacteria belonging at least to P. autumnale lineages has ancient origins. It exceeded the breakup of Gondwana and originated somewhere on the supercontinent before 442 Ma and 297 Ma according to estimates by R8S and beast, respectively (Fig. 4). Former estimates suggest the coevolution of Phormidium as a representative of simple filamentous Oscillatoriales with the emergence of vascular plants in the Ordovic–Silur transition and their progress of the colonization of land (Willis & McElwain, 2002) where the mats could serve as further soil stabilizers (Hodkinson et al., 2003). Latter beast estimates are more conservative; however, it connects the evolution of Phormidium to carbon–perm transition and to other events in Earth's history. The evolution of flying insects (Peart & Roberts, 2008) as Trichoptera (Shear & Kukalova-Peck, 1990) or Ephemeroptera (Knecht et al., 2011) feeding on algal mats in the Late Carboniferous could enhance the dispersibility of freshwater cyanobacteria. Our data indicated the rapid acclimatization of Phormidium to harsh climate of Antarctic glaciations after the onset of glaciation of the continent in the Eocene (Fig. 4). They confirmed the hypothesis of organism surviving through ice ages and the existence of refuges within Antarctica. The morphology of strains suggested the coevolution of lineages and formation of complex associations with different morphologies (Figs 1 and 2), which could possibly exploit, to a large extent, the various biotopes of the poor environment of Antarctica. Although we lacked sufficient amount of data from locations other than the Antarctic Peninsula to confirm this hypothesis for the whole continent, we assume that this hypothesis is universally valid throughout Antarctica.
We would like to thank the Ministry of Education of the Czech Republic (Kontakt ME 934, ME 945, INGO LA 341 and LM – 2010009 CzechPolar) for funding our research. We are very grateful to Dr Alena Lukešová (Biology Centre, Institute of Soil Biology, České Budějovice) and Dr Ĺubomír Kováčik (Comenius University, Department of Botany, Bratislava) for the strains they provided for our study. In addition, we are also very grateful for the technical assistance provided by Mrs Jana Šnokhousová and Mrs Dana Švehlová.