Origin of marine planktonic cyanobacteria

Phylogenetic relationships

An increase in genome sequencing and taxon-sampling have allowed for broad coverage of a range of morphologies, lifestyles, and metabolisms within cyanobacteria²². The analyses performed here included a large phylogenetic data set consisting of 131 genome taxa with a total of 56,251 amino acids (aa) and 4,555 base pairs (bp). Whilst analyses have recovered well-supported monophyletic groups previously reported⁹,²²,²³,²⁴,²⁵,²⁶, new genomic data have revealed novel deep-branching relationships of major cyanobacteria lineages²²,²⁴ (Fig. 1 and Supplementary Fig. S1). In this study Pseudanabaena appears as an early divergent lineage within cyanobacteria (Fig. 1 and Supplementary Fig. S1) occurring in 88% of the Maximum Likelihood trees generated for each gene alignment (137 genes) generated in SATé 2.2.3²⁷. A basal position for Pseudanabaena is consistent with recent large-scale multi gene studies²²,²⁸. Previous studies suggesting that Pseudanabaena is a derived lineage were based on SSU rRNA datasets⁹,²⁵.

Genomic data have also clarified problematic phylogenetic relationships such as the positioning of the filamentous LPP group (Fig. 1 and Supplementary Fig. S1). New data strongly support sister relationships between the LPP clade (Supplementary Fig. S1) and Prochlorothrix, Synechoccocous elongatus and the SynPro clade (Synechococcus, Prochlorococcus, Cyanobium). While the inclusion of recently sequenced genomes²²,²⁴,²⁹,³⁰ suggest a new placement for Trichodesmium (Fig. 1 and Supplementary Fig. S1), more Oscillatoria-like genomes are needed to fully understand the placement of this important lineage. Modern marine planktonic cyanobacteria evolved within two major groups of cyanobacteria, here referred to as the Microcyanobacteria and the Macrocyanobacteria since they are well-supported monophyletic clades (Supplementary Fig. S1). Whilst the Microcyanobacteria contain lineages with smaller cell diameters (<3 μm), the Macrocyanobacteria contain lineages with larger cell diameters (>3 μm; Supplementary Fig. S3). The Macrocyanobacteria are the most taxonomically and ecologically diverse clade including lineages such as Synechocystis, Pleurocapsa, Microcystis, Trichodesmium and the Nostocales, amongst others (Fig. 1).

Relaxed molecular clock analyses

Age divergences were estimated using two independent data sets, RNA (SSU and LSU: 4,555 bp) and proteins (18 genes: 4,980 aa), and applying a Bayesian approach³¹,³². Four calibration points were implemented, three of which have been previously used⁹,²⁵. Relaxed molecular clock analyses were performed under the independent-rates model³³, which has been shown to be the best fitting molecular clock model for cyanobacteria based on Bayes Factors⁹. Four different models of molecular evolution were implemented for proteins and RNA in MCMCtree and the CAT- GTR model for proteins and RNA in Phylobayes (Table 1). The implementation of two different maximum ages for the origin of oxygenic photosynthesis (i.e., 3,000 and 2,700 Myr) resulted in different age estimates for the origin of filamentous forms (node 2). While an older maximum age (3,000 Myr) predicts the origin of filamentous forms (node 2) before the GOE with estimates ranging between 2,665 and 2,559 Mya, a younger maximum age (2,700 Myr) predicts filamentous forms appearing around the time of the GOE between 2,460 and 2,351 Mya (Fig. 1, Table 1). Overall an older maximum age (3,000 Myr) tends to make ages older across all analyses.

Results were consistent across models of molecular evolution within each data set. There is strong evidence for a Neoproterozoic or early Cambrian origin for marine unicellular N-fixers (i.e., the Crocosphaera clade) and the filamentous Nodularia spumigena CCY9414. Age estimates appear to be younger for Prochlorococcus (nodes 9) and Synechococcus (node 10) based on the nucleotide data set, in contrast to the protein data set (Table 1). All analyses however provide robust evidence for the relatively late evolution of marine planktonic cyanobacteria. Other marine N-fixers evolved during the Phanerozoic such as Richelia (a diatom symbiont) and the cyanobacterium UCYN-A clade (a coccolithophore symbiont; Fig. 1). Age estimates across all analyses are summarized in Table 1 and Fig. 1, and are mostly in broad agreement.

Bayesian trait evolution analyses

The earliest cyanobacteria were likely unicellular (node 1) and inhabited low salinity environments (Fig. 2 and Supplementary Fig. S4). Living relatives of these early divergent lineages have been isolated from terrestrial/freshwater environments (e.g., Pseudanabaena PCC6802 Cyanothece PCC7425, Fig. 1), hot-springs (e.g., Thermosynechococcus elongatus BP-1) and coastal marine habitats (e.g., Acaryochloris and Synechococcus PCC7336). Bayesian stochastic character mapping analyses revealed that filamentous cyanobacteria evolved early on and different molecular clock analyses indicate filamentous forms evolved around 2,665 to 2,351 Mya and the GOE (node 2; Figs 1 and Supplementary Fig. S2). Ancestors of early filamentous forms likely resembled modern relatives of Pseudanabaena and the LPP clade (nodes 2 and 3; Fig. 2). All Basal Lineages and the Microcyanobacteria have retained small cell diameters exhibiting cells that are less than 3 μm, with most lineages exhibiting diameters that are less than 2 μm (Supplementary Fig. S3 and Supplementary Table S2). Interestingly, further decrease in cell diameter characterizes the evolution of the marine Prochlorococcus within the SynPro clade²⁶. Also a switch from filamentous to unicellular cell types occurred (node 5; Figs 1 and 2) around 1,994 to 1,421 Mya (Table 1).

All analyses suggest that the Macrocyanobacteria clade, exhibiting larger cell diameters (>3 up to 50 μm), may have evolved just after the GOE with age estimates ranging between 2,386 and 1,894 Mya (node 4; Fig. 1 and Table 1 and Supplementary Fig. S3). Within this clade two opposite evolutionary trends were found: 1) an increase in cell diameter (e.g., Fischerella and Mastigocladopsis) within the Nostocales, and 2) a decrease in cell diameter (node 7) within the clade containing Microcystis and Crocosphaera relatives (Supplementary Table S2, Supplementary Fig. S3). A switch from filamentous to unicellular forms also occurred (node 7, Fig. 1 and Supplementary Fig. S2) around 1,437 to 1,047 Mya in freshwater habitats (Supplementary Fig. S4). Whilst unicellular marine N-fixing cyanobacteria (e.g., Crocosphaera and relatives) and Nodularia spumigena CCY9414 diverged from freshwater ancestors, Trichodesmium evolved from filamentous coastal marine lineages (Fig. 1 and Supplementary Fig. S4).

Stem vs crown groups

Recent genomic data have improved the resolution of the cyanobacteria tree of life helping with the interpretation of the geological record⁹,²⁵,²⁸. Cyanobacteria fossils with a cell diameter bigger than 3 μm appeared in the Belcher Subgroup with fossils such as colonial coccoids (Eoentophysalis) and colonial ellipsoids (Eosynechococcus)³⁴,³⁵. Oscillatoria-like filamentous fossils (e.g., Gunflintia) appeared in the Gunflint iron formation³⁶, and Halythrix sp. in the Belcher subgroup³⁴. At approximately 1,900 Myr, microfossils with increased cell diameters as well as sheaths became common³⁵. These findings are consistent with the evolutionary studies shown here in which ancestors with inferred cell diameters larger than 3 μm (node 4), the Macrocyanobacteria, postdate the GOE (Table 1, Fig. 2)²⁵. It is therefore likely that the first appearance of reliable cyanobacteria fossils observed at around 2,000 Myr³⁵,³⁷ is indicative of an ancient cyanobacteria radiation²³. Interestingly, age estimates based on molecular clock studies show a lag between the early origin of oxygenic photosynthesis²⁵,²⁸,³⁸,³⁹ and the first reliable evidence of cyanobacteria in the fossil record at around 2,000 Myr³⁵,³⁷,⁴⁰. The Macrocyanobacteria clade (node 4) also evolved traits necessary for establishing thick laminated mats²³,²⁵,²⁶ and have shown a significant shift in diversification rates³⁹. It is not surprising that this clade contains the highest taxonomic and ecological diversity of modern cyanobacteria.

Age estimates shown here suggests that there was a delay between the appearance of the first reliable cyanobacteria fossils and the ancestors of the crown groups containing marine planktonic cyanobacteria (e.g., Nostocales/Gloeocapsa, Arthrospira/Trichodesmium and Pleurocapsa/Microcystis/Crocosphaera clades; Table 1; Fig. 1). The great majority of modern cyanobacteria can be traced back to the late Paleoproterozoic and the Mesoproterozoic, implying that older cyanobacteria fossils (e.g., Gunflintia) belong to stem groups with no living relatives. Crown groups with morphologies that required cell differentiation and division of labor (e.g., Nostocales) evolved during the Mesoproterozoic (Fig. 1). Cell differentiation is particularly important for some marine N-fixing planktonic cyanobacteria that evolved during the Neoproterozoic (e.g., Nodularia spumigena) and Cretaceous (e.g., Richelia). The evolution of cell differentiation mechanisms would have involved a specific program of gene expression including the induction of regulatory genes and of genes encoding the proteins for the morphological and biochemical differentiation for specialized cells (i.e., heterocyst)²⁶. Comparative genomic studies have shown that more complex morphologies characteristic of crown groups (e.g., Nostocales) required the evolution of additional genes involved in signal transduction and transcription-related functional categories²². Genomic and trait evolution studies have also revealed that more complex morphologies within cyanobacteria exhibit bigger genome sizes²⁶ presumably as a result of the more elaborate metabolic processes involved in these lineages⁴¹.

A historical perspective

Previous broad taxonomic and phylogenomic studies of prokaryotes studies have inferred a terrestrial/freshwater ancestry of cyanobacteria⁴²,⁴³. Trait evolution studies of cyanobacteria, implementing large genomic data sets studies, have come to similar conclusions²³,²⁵,²⁹. The Bayesian stochastic character mapping analyses presented here confirm that cyanobacteria first evolved in freshwaters. Moreover, most cyanobacterial lineages inhabited benthic, terrestrial and/or shelf environments for most of the Proterozoic. Interestingly, in modern habitats, benthic cyanobacteria are much more taxonomically diverse. This is perhaps due to the great variety of available niches in coastal environments (e.g., intertidal or infralittoral areas)⁴⁴. The early establishment of mat-forming filamentous cyanobacteria (Fig. 1)²⁵,²⁸ and subsequent dominance of benthic microbial communities would have restricted primary productivity to terrestrial and ocean margins (Fig. 2). The small area of fresh waters and ocean margins imply that the global biogeochemical impact of oxygenic photosynthesis would have been minimal until cyanobacteria started colonizing the open ocean⁹,²⁵, which currently covers approximately two-thirds of the Earth’s surface⁴¹. Some marine planktonic lineages had a freshwater ancestor (Fig. 2). This is illustrated by the unicellular marine clades such as Crocosphaera and SynPro. Marine lineages adapted to marine environments by acquiring the machinery that enables them to osmoregulate in marine environments such as the set of genes responsible for the synthesis of compatible solutes: glucosylglycerol (GG), glucosylglycerate (GGA) and glycine betaine (GB)⁴⁵,⁴⁶.

This study has revealed that decrease in cell size and a switch from filamentous to unicellular forms or loss of filamentous forms, in the lead up to the origin of the Crocosphaera and the SynPro clades, likely played a key role in the emergence of a planktonic life style. Convergent evolution with regard to the emergence of unicellular phytoplankton forms suggests similar selective pressures taking place (e.g., nutrient starvation) on the evolutionary history of these lineages. Other strategies such as gas vesicles evolved to cope with buoyancy regulation amongst some marine filamentous lineages such as Trichodesmium and Nodularia spumigena.

Marine planktonic habitats likely provided a challenging environment for cyanobacteria to proliferate into since the ocean remained anoxic for most of the Proterozoic⁴,⁴⁷,⁴⁸. Under anoxic conditions, including episodes of euxinia (anoxic environments with the presence of hydrogen sulphide) and ferruginous conditions (see recent reviews on ocean geochemistry of the Pre-Cambrian⁴,¹¹), key trace metals essential for N-fixation would have been depleted. This is particularly the case for micronutrients such as molybdenum (Mo) an essential constituent of the nitrogenase enzyme involved in N-fixation⁴⁹. While marine mat-forming cyanobacteria such as Microcoleus chthonoplastes are capable of performing sulphide-dependent anoxygenic photosynthesis, this biogeochemical process appears to serve as a detoxification mechanism⁵⁰ in response to an inherent active sulphur cycle found in microbial mats. In microbial mats, the establishment of vertically stratified layers (known as laminated structures) allows for the spatial separation of oxygenic photosynthesis and N-fixation (an anaerobic process inhibited by oxygen)⁵⁰. The segregation of biogeochemical processes could have persisted through most of the Pre-Cambrian, allowing the coexistence of N-fixers and oxygenic phototrophs. Interestingly, Trichodesmium is the only modern planktonic filamentous N-fixer that diverged from mat-forming relatives⁵¹.

Marine planktonic cyanobacteria

Only a few lineages within cyanobacteria adapted to the lack of nutrients (oligotrophy) characteristic of the Earth’s open ocean. Age estimates suggest that there was an interval of more than a billion years between the timing of origin of the common ancestor of the Macrocyanobacteria and the origin of modern marine planktonic N-fixers (Fig. 1, Table 1). Nitrogen-fixing cyanobacteria evolved different morphologies and physiological strategies within the Macrocyanobacteria. At least three lineages evolved independently towards the end of the Pre-Cambrian: Crocosphaera and relatives, Trichodesmium, and Nodularia spumigena. Ancestors of the unicellular N-fixers, Crocosphaera clade, underwent decrease in cell diameters and a switch from filamentous to unicellular forms. While unicellular marine N-fixers, the Crocosphaera clade, diverged from freshwater relatives, Trichodesmium evolved from filamentous mat-forming cyanobacteria which have modern relatives found in benthic and in marine littorals⁹,⁵¹. Within the Nostocales, Nodularia spumigena CCY9414, a lineage currently found in the Baltic Sea in salty or brackish waters, diverged from freshwater relatives towards the end of the Pre-Cambrian (Fig. 1). Recent planktonic lineages such as Richelia and the cyanobacterium UCYN-A clade evolved as symbionts during the Cretaceous.

Decrease in cell diameter was part of the evolutionary history of the major primary producers, Prochlorococcus and Synechococcus within the Microcyanobacteria (Fig. 1). The abundant marine Prochlorococcus and Synechococcus shared a common ancestor and evolved within the Microcyanobacteria (Fig. 1). The SynPro clade is sister to the filamentous Prochlorothrix and nested within the LPP clade (Fig. 1 and Supplementary Fig. S1). This phylogenetic relationship indicates that there was a switch from filamentous to unicellular forms³⁹,⁵² in which cell adhesion and intracellular communication were likely lost. Within the evolution of the marine SynPro clade, there is also a trend in the decrease of genome and cell size¹⁷,²⁶. Convergent evolution with regard to a decrease in cell diameter highlights the advantage of smaller cells when inhabiting oligotrophic environments. In modern oceans small phytoplankton cells usually dominate phytoplankton communities under oligotrophic conditions, such as the oceanic gyres, whereas larger phytoplankton cells are more abundant along continental margins and in upwelling zones, where nutrient concentrations tend to be higher and more variable⁵³. Divergence times (Table 1) mostly agree with a Neoproterozoic or early Phanerozoic origin for Prochlorococcus (node 9) and the marine Synechococcus (node 10).

Lack of evidence for early marine planktonic cyanobacteria

After the GOE, it has been assumed that marine cyanobacteria were responsible for the increase in primary productivity (Lomagundi–Jatuli excursion) and carbon burial around 2,200 to 2,060 Mya²,⁵⁴,⁵⁵. Yet if marine planktonic cyanobacteria were involved in the carbon burial (e.g., Lomagundi-Jatuli excursion), evolutionary studies have provided no evidence so far supporting the survival of early marine planktonic lineages (Fig. 1 and Supplementary Fig. S4) or perhaps these lineages have not yet been discovered. While marine living planktonic cyanobacteria cannot be traced back to the early Paleoproterozoic (Figs 1 and 2), we cannot discard the possibility that early marine cyanobacteria lineages went extinct after the GOE due to changes in marine water chemistry resulting from freely available oxygen (e.g., euxinic conditions)⁴,¹¹,¹². It is clear that more efforts are needed studying early divergent lineages of cyanobacteria in order to unravel the transition from a terrestrial to marine biosphere.

Alignment and taxon sampling

Alignments including 135 protein-coding genes (56,251 aa) and two ribosomal RNAs (4,555 bp) were analysed for 131 genome taxa. All sequence data for the 131 cyanobacterial genomes were obtained from GenBank (http://www.ncbi.nlm.nih.gov) and using Geneious R6. The chosen genes (135 proteins and two ribosomal RNA: SSU and LSU) were universally present in cyanobacterial taxa, evolutionarily conserved and had a minimum number of gene duplications²³,⁵⁶. Principal coordinates analyses were used to identify ortholog genes that belong to a conserved ‘core’ gene set in cyanobacteria⁵⁶. For the protein data set, genes chosen represent a wide diversity of cellular functions. These functions range from Information Processing (IP; transcription, translation, DNA replication and repair), Metabolism (Met), Cellular Processes (CP; including cell division, cell envelope biogenesis, motility and secretion) and General Function Prediction (GFP). For a detailed list with names and description of the genes included in this study see Blank and Sánchez-Baracaldo²⁵. Each gene was aligned independently using SATé 2.2.3²⁷, a multiple sequence alignment and phylogenetic reconstruction program. Single gene alignments generated in SATé were imported into Mesquite v. 2.75⁵⁷ to obtain ‘nexus’ and ‘phylip’ format files for subsequent analyses. Single alignments were later concatenated into a single nexus format file using Sequence Matrix v 100.0⁵⁸. Two concatenated matrices were obtained: one for protein-coding genes and a second one for ribosomal RNAs; both of these matrices were used to estimate tree topologies as described below.

Phylogenetic analyses

Maximum likelihood analyses were performed in RAxML 7.2.6⁵⁹ and Phylobayes³². A multiple gene approach was implemented to establish the deep-branching relationships in the cyanobacterial tree. All sequence data for 131 cyanobacterial genomes were obtained from GenBank (http://www.ncbi.nlm.nih.gov) and using Geneious 6.1.4. A total of 135 protein-coding genes (with 56,251 aa) and two ribosomal RNAs (with a total of 4,555 bp) were used to establish phylogenetic relationships of cyanobacteria. ProTest v.2.4⁶⁰ was used to estimate the best model of evolution for the protein set. To analyse the protein sequences I implemented the LG model and G (gamma-distribution with 4 rate categories). Matrices containing the protein and RNA data set were imported into RAxML GUI v.1.1⁶¹ and up to 50 maximum likelihood trees were generated. To obtain statistical support analyses were performed in RAxML 7.0.3⁵⁹ (Supplementary Fig. S1). RAxML analyses recovered the same well-supported monophyletic groups previous reported by recent phylogenomic studies²²,²³,²⁵,²⁸,⁶².

Bayesian divergence time estimation

Divergence times were estimated implementing a Bayesian relaxed molecular clock approach in MCMCTree³¹ and Phylobayes³² (Table 1). Two independent data sets were assembled to estimate age divergences: 1) eighteen proteins, and 2) two RNA genes (SSU and LSU). The independent-rates model was implemented as previous studies of cyanobacteria have shown that this model is favored over the auto-correlated rates model⁹. Since cyanobacteria have an ancient origin, independent rates more likely represent greater variation in inherited factors in contrast to auto-correlated rates⁹. Because the current implementation of MCMCtree and Phylobayes do not allow the use of mixed (nucleotides and amino acids) data sets. Ages were estimated using both type of data sets separately as follows: 1) Protein data set with a total of 18 proteins (AtpA, AtpB, AtpH, L1, L4, L5, NdhH, PetB, PetD, PsaA, PsbE, RbcL, S10, S13, S19, SecY, TufA and Ycf3), and 2) nucleotide data set including two RNAs (SSU and LSU). Supplementary Table S1 contains a list of gene names and substitution rates for the analyses performed in this study.

In MCMCtree, I implemented four models of nucleotide and protein evolution (See Supplementary Information on line). For all age calibrations, soft bounds were specified with 2.5% tail probabilities above/below these limits, allowing for molecular data to correct for conflicting fossil information. In Phylobayes, I implemented the CAT-GTR replacement model for both nucleotides and amino acids³². For all non-calibrated nodes, I used a birth-death prior on divergence times I also performed two separate experiments two independent experiments with permissive gamma distributed root priors that allowed a 95% credibility interval of the root node to range between 2,320–2,700 Mya and 2,320–3,000 Mya. See supplementary information for a more detailed description of the relaxed molecular clock analyses performed.

Fossil constraints

Experiments reported here (Table 1) were performed implementing two maximum ages for the cyanobacterial root: 2,700 Myr⁶³ and 3,000 Myr⁶⁴. The minimum age for the cyanobacterial root was set at 2,320 Myr (the rise in atmospheric oxygen)². Based on geochemical evidence a younger (2,500 Myr)⁶⁵ age is also reported for the origin of oxygenic photosynthesis. I also calibrated the trees with fossils exhibiting unique morphological features that could be assigned to well-supported groups such as the Nostocales and the Pleurocapsales. In the Nostocales, akinetes are thick-walled dormant cells to enable a response to extreme cold and desiccation resistant environmental conditions⁶⁶. These specialized cells are present in most species amongst the Nostocales and likely evolved once in this group. Akinetes have been shown to have a single ancestor based on trait evolution studies²³. A maximum age of 2,100 Myr was used for this monophyletic group for which it has been hypothesised that specialised cells such as the heterocyst evolved in response to free available oxygen in the atmosphere⁶⁷. The Pleurocapsales are characterised by having multiple fission, a unique phenotypic property that distinguishes members of this group from other cyanobacteria. The Pleurocapsales also belong to a well-supported monophyletic group (Supplementary Fig. S1) including two strains of Pleurocapsa (PCC 7319 and PCC 7327). A minimum age of 1,700 Myr⁶⁸ and a maximum age of 1,900 Myr (Gunflint iron formation and the first appearance of reliable cyanobacteria fossils observed in the fossil record³⁵,³⁷ were implemented. Finally, a maximum age of 110 Myr was implemented for Hemiaulus⁶⁹ as these organisms host the symbiont Richelia¹⁹,²⁰,²¹.

Bayesian inference of character evolution

To infer the evolution of cell type, cell diameter and habitat, I used Bayesian stochastic character mapping⁷⁰. Analyses were implemented in SIMMAP v1.5⁷¹ and used relative time calibrated trees generated in MCMCtree³¹ for the protein set as described above. No prior on the rate parameter was used, as I wanted to use the branch lengths as a direct estimate of rate of evolution. A β prior is implemented in SIMMAP on the symmetry of the transition rate matrix. Cell type or morphology were coded as 0 = unicellular, and 1 = filamentous. Cell diameter data were coded as discrete characters, where 0 = average cell diameter raging from 1 to 2 μm, 1 = average cell diameter raging 2 to 3 μm, 2 = average cell diameter raging 3 to 5 μm, and 3 = average cell diameter equal to or greater than 5 μm. Habitat were coded as 0 = freshwater, and 1 = brackish, marine or hypersaline. Character states were obtained from the Bergey’s Manual of Systematic Bacteriology⁷², previous studies of trait evolution of cyanobacteria²⁵, and other cyanobacteria studies²². For binary characters (habitat and cell type) the bias parameter was drawn from a symmetrical β prior. Since SIMMAP uses a symmetrical β prior on the symmetry of the transition rate matrix, this influences the degree to which transitions favor state 0 over 1. The shape of the β distribution is described by the α parameter and discretized into κ categories. I performed sensitivity experimenting using three different α distributions, where α used the following values, 0.1, 1 and 10. Analyses for habitat and cell type used α = 1. For multi-state characters (cell diameter) the bias parameter between states is specified as simply 1/κ, where k is the number of states. The overall rate of substitution for both of these classes is a branch length multiplier drawn from a prior gamma distribution. Supplementary Table S2 contains the characters states used for the three characters studied here.

Supplementary Information

1. Holland H. D. The oxygenation of the atmosphere and oceans. Philos. T. R. Soc. B. 361, 903–915, doi: 10.1098/rstb.2006.1838 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
2. Bekker A. et al. Dating the rise of atmospheric oxygen. Nature 427, 117–120, doi: 10.1038/nature02260 (2004). [DOI] [PubMed] [Google Scholar]
3. Planavsky N. J. et al. Low Mid-Proterozoic atmospheric oxygen levels and the delayed rise of animals. Science 346, 635–638 (2014). [DOI] [PubMed] [Google Scholar]
4. Lyons T. W., Reinhard C. T. & Planavsky N. J. The rise of oxygen in Earth’s early ocean and atmosphere. Nature 506, 307–315 (2014). [DOI] [PubMed] [Google Scholar]
5. Sahoo S. K. et al. Ocean oxygenation in the wake of the Marinoan glaciation. Nature. 489, 546–549, doi: 10.1038/nature11445 (2012). [DOI] [PubMed] [Google Scholar]
6. Och L. M. & Shields-Zhou G. A. The Neoproterozoic oxygenation event: Environmental perturbations and biogeochemical cycling. Earth-Sci Rev. 110, 26–57, doi: 10.1016/j.earscirev.2011.09.004 (2012). [DOI] [Google Scholar]
7. Canfield D. E., Poulton S. W. & Narbonne G. M. Late-Neoproterozoic Deep-Ocean Oxygenation and the Rise of Animal Life. Science 315, 92–95 (2007). [DOI] [PubMed] [Google Scholar]
8. Fike D. A., Grotzinger J. P., Pratt L. M. & Summons R. E. Oxidation of the Ediacaran Ocean. Nature 444, 744–747. [DOI] [PubMed] [Google Scholar]
9. Sánchez-Baracaldo P., Ridgwell A. & Raven J. A. A Neoproterozoic transition in the marine nitrogen cycle. Curr. Biol. 24, 652–657, doi: 10.1016/j.cub.2014.01.041 (2014). [DOI] [PubMed] [Google Scholar]
10. Erwin D. H. et al. The Cambrian Conundrum: Early Divergence and Later Ecological Success in the Early History of Animals. Science 334, 1091–1097, doi: 10.1126/science.1206375 (2011). [DOI] [PubMed] [Google Scholar]
11. Poulton S. W. & Canfield D. E. Ferruginous Conditions: A Dominant Feature of the Ocean through Earth’s History. Elements 7, 107–112, doi: 10.2113/gselements.7.2.107 (2011). [DOI] [Google Scholar]
12. Anbar A. D. & Knoll A. H. Proterozoic ocean chemistry and evolution: a bioinorganic bridge? Science 297, 1137–1142, doi: 10.1126/science.1069651 (2002). [DOI] [PubMed] [Google Scholar]
13. Field C. B., Behrenfeld M. J., Randerson J. T. & Falkowski P. Primary Production of the Biosphere: Integrating Terrestrial and Oceanic Components. Science 281, 237–240 (1998). [DOI] [PubMed] [Google Scholar]
14. Zehr J. P. Nitrogen fixation by marine cyanobacteria. Trends Microbiol. 19, 162–173 doi: 10.1016/j.tim.2010.12.004 (2011). [DOI] [PubMed] [Google Scholar]
15. Thompson A. W. et al. Unicellular cyanobacterium symbiotic with a single-celled eukaryotic alga. Science 337, 1546–1550, doi: 10.1126/science.1222700 (2012). [DOI] [PubMed] [Google Scholar]
16. Johnson Z. I. et al. Niche partitioning among Prochlorococcus ecotypes along ocean-scale environmental gradients. Science 311, 1737–1740 (2006). [DOI] [PubMed] [Google Scholar]
17. Scanlan D. J. et al. Ecological Genomics of Marine Picocyanobacteria. Microbiol. Mol. Biol. Rev. 73, 249–299 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
18. Flombaum P. et al. Present and future global distributions of the marine Cyanobacteria Prochlorococcus and Synechococcus. Proc. Natl. Acad. Sci. USA 110, 9824–9829, doi: 10.1073/pnas.1307701110 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
19. Foster R. A. et al. Nitrogen fixation and transfer in open ocean diatom-cyanobacterial symbioses. The ISME J. 5, 1484–1493, doi: 10.1038/ismej.2011.26 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
20. Villareal T. Laboratory culture and preliminary characterization of the nitrogen-fixing Rhizosolenia-Richelia symbiosis. Mar. Ecol. 11, 117–132 (1990). [Google Scholar]
21. Janson S., Wouters J., Bergman B. & Carpenter E. J. Host specificity in the Richelia-diatom symbiosis revealed by hetR gene sequence analysis. Environ. Microbiol. 1, 431–438 (1999). [DOI] [PubMed] [Google Scholar]
22. Shih P. M. et al. Improving the coverage of the cyanobacterial phylum using diversity-driven genome sequencing. P. roc Natl. Acad. Sci. USA 110, 1053–1058, doi: 10.1073/pnas.1217107110 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
23. Sánchez-Baracaldo P., Hayes P. K. & Blank C. E. Morphological and habitat evolution in the Cyanobacteria using a compartmentalization approach. Geobiology 3, 145–165 (2005). [Google Scholar]
24. Bombar D., Heller P., Sánchez-Baracaldo P., Carter B. J. & Zehr J. P. Comparative genomics reveals surprising divergence of two closely related strains of uncultivated UCYN-A cyanobacteria. The ISME J. 8, 2530–2542, doi: 10.1038/ismej.2014.167 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
25. Blank C. E. & Sánchez-Baracaldo P. Timing of morphological and ecological innovations in the cyanobacteria–a key to understanding the rise in atmospheric oxygen. Geobiology 8, 1–23, doi: 10.1111/j.1472-4669.2009.00220.x (2010). [DOI] [PubMed] [Google Scholar]
26. Larsson J., Nylander J. A. & Bergman B. Genome fluctuations in cyanobacteria reflect evolutionary, developmental and adaptive traits. BMC Evol. Biol. 11, 187 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
27. Liu K., Raghavan S., Nelesen S., Linder C. R. & Warnow T. Rapid and accurate large-scale coestimation of sequence alignments and phylogenetic trees. Science 324, 1561–1564, doi: 10.1126/science.1171243 (2009). [DOI] [PubMed] [Google Scholar]
28. Schirrmeister B. E., Gugger M. & Donoghue P. C. J. Cyanobacteria and the Great Oxidation Event: evidence from genes and fossils. Palaeontology, doi: 10.1111/pala.12178 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
29. Dagan T. et al. Genomes of Stigonematalean Cyanobacteria (Subsection V) and the Evolution of Oxygenic Photosynthesis from Prokaryotes to Plastids. Genome Biol. Evol. 5, 31–44 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
30. Donia M. S. et al. Complex microbiome underlying secondary and primary metabolism in the tunicate-Prochloron symbiosis. Proc. Natl. Acad. Sci. USA 108, E1423–E1432 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
31. Yang Z. PAML 4: Phylogenetic Analysis by Maximum Likelihood. Mol. Biol. Evol. 24, 1586–1591 (2007). [DOI] [PubMed] [Google Scholar]
32. Lartillot N., Lepage T. & Blanquart S. PhyloBayes 3: a Bayesian software package for phylogenetic reconstruction and molecular dating. Bioinformatics 25, 2286–2288. [DOI] [PubMed] [Google Scholar]
33. Drummond A. J., Ho S. Y. W., Phillips M. J. & Rambaut A. Relaxed Phylogenetics and Dating with Confidence. Plos Biol. 4, e88 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
34. Hofmann H. J. Precambrian microflora, Belcher Islands, Canada: significance and systematics. J. Paleontol. 50, 1040–1073 (1976). [Google Scholar]
35. Golubic S. & Seong-Joo L. Early cyanobacterial fossil record: preservation, palaeoenvironments and identification. Eur. J. Phycol. 34, 339–348 (1999). [Google Scholar]
36. Wacey D. et al. Nanoscale analysis of pyritized microfossils reveals differential heterotrophic consumption in the ~1.9-Ga Gunflint chert. Proc. Natl. Acad. Sci. USA 110, 8020–8024, doi: 10.1073/pnas.1221965110 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
37. Sergeev V. N., Gerasimenko L. M. & Zavarzin G. A. Proterozoic history and present state of cyanobacteria. Mikrobiologiia 71, 725–740 (2002). [PubMed] [Google Scholar]
38. Falcón L. I., Magallón S. & Castillo A. Dating the cyanobacterial ancestor of the chloroplast. The ISME J., 4, 777–783, doi: 10.1038/ismej.2010.2 (2010). [DOI] [PubMed] [Google Scholar]
39. Schirrmeister B. E., de Vos J. M., Antonelli A. & C. B. H. Evolution of multicellularity coincided with increased diversification of cyanobacteria and the Great Oxidation Event. Proc. Natl. Acad. Sci. USA 110, 1791–1796 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
40. Brasier M. D. Green algae (Chlorophyta) and the question of freshwater symbiogenesis in the early proterozoic. J. Phycol. 49, 1036–1039 (2013). [DOI] [PubMed] [Google Scholar]
41. Raven J. A., Beardall J., Larkum A. W. & Sánchez-Baracaldo P. Interactions of photosynthesis with genome size and function. Philos. T. R. Soc. B. 368, 20120264, doi: 10.1098/rstb.2012.0264 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
42. Battistuzzi F. U., Feijao A. & Hedges S. B. A genomic timescale of prokaryote evolution: insights into the origin of methanogenesis, phototrophy, and the colonization of land. BMC Evol. Biol. 4, 44, doi: 10.1186/1471-2148-4-44 (2004). [DOI] [PMC free article] [PubMed] [Google Scholar]
43. Sleep N. H. & Bird D. K. Evolutionary ecology during the rise of dioxygen in the Earth’s atmosphere. Philos. T. R. Soc. B. 363, 2651–2664, doi: 10.1098/rstb.2008.0018 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
44. Coelho S. M., Simon N., Ahmed S., Cock J. M. & Partensky F. Ecological and evolutionary genomics of marine photosynthetic organisms. Mol. Ecol. 22, 867–907, doi: 10.1111/mec.12000 (2012). [DOI] [PubMed] [Google Scholar]
45. Hagemann M. Genomics of Salt Acclimation: Synthesis of Compatible Solutes among Cyanobacteria. Adv. Bot. Res. 65, 27–55, doi: 10.1016/B978-0-12-394313-2.00002-0 (2013). [DOI] [Google Scholar]
46. Hagemann M. Molecular biology of cyanobacterial salt acclimation. FEMS Microbiol. Rev. 35, 87–123, doi: 10.1111/j.1574-6976.2010.00234.x (2011). [DOI] [PubMed] [Google Scholar]
47. Scott C. et al. Tracing the stepwise oxygenation of the Proterozoic ocean. Nature 452, 456–U455, doi: 10.1038/nature06811 (2008). [DOI] [PubMed] [Google Scholar]
48. Planavsky N. J. et al. Widespread iron-rich conditions in the mid-Proterozoic ocean. Nature 477, 448–U495, doi: 10.1038/nature10327 (2011). [DOI] [PubMed] [Google Scholar]
49. Zerkle A. L., House C. H., Cox R. P. & Canfield D. E. Metal limitation of cyanobacterial N2 fixation and implications for the Precambrian nitrogen cycle. Geobiology 4, 285–297 (2006). [Google Scholar]
50. Stal L. Cyanobacterial mats and stromatolites. In: Ecology of Cyanobacteria II. Springer. (Netherlands), pp 65–126. Whitton B. A. (Springer, 2012). [Google Scholar]
51. Abed R. M. M., Palinska K. A., Camoin G. & Golubic S. Common evolutionary origin of planktonic and benthic nitrogen-fixing oscillatoriacean cyanobacteria from tropical oceans. FEMS Microbiol. Lett. 260, 171–177, doi: 10.1111/j.1574-6968.2006.00321.x (2006). [DOI] [PubMed] [Google Scholar]
52. Schirrmeister B. E., Antonelli A. & Bagheri H. C. The origin of multicellularity in cyanobacteria. BMC Evol. Biol. 11, doi: Artn 4510.1186/1471-2148-11-45 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
53. Flombaum P. & et al. Present and future global distributions of the marine Cyanobacteria Prochlorococcus and Synechococcus. Proc. Natl. Acad. Sci. USA 110, 9824–9829, doi: 10.1073/pnas.1307701110 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
54. Papineau D., Purohit R., Fogel M. L. & Shields-Zhou G. A. High phosphate availability as a possible cause for massive cyanobacterial production of oxygen in the Paleoproterozoic atmosphere. Earth Planet Sc. Lett. 362, 225–236, doi: 10.1016/j.epsl.2012.11.050 (2013). [DOI] [Google Scholar]
55. Kump L. R. et al. Isotopic Evidence for Massive Oxidation of Organic Matter Following the Great Oxidation Event. Science 334, 1694–1696, doi: 10.1126/science.1213999 (2011). [DOI] [PubMed] [Google Scholar]
56. Shi T. & Falkowski P. G. Genome evolution in cyanobacteria: the stable core and the variable shell. Proc. Natl. Acad. Sci. USA 105, 2510–2515, doi: 10.1073/pnas.0711165105 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
57. Maddison W. & Maddison D. Mesquite: a modular system for evolutionary analysis. Version 2.75. Available from http://mesquiteproject.org (2009). (Accessed: 10th January 2012).
58. Vaidya G., Lohman D. J. & Meier R. SequenceMatrix: concatenation software for the fast assembly of multi-gene datasets with character set and codon information. Cladistics 27, 171–180, doi: 10.1111/j.1096-0031.2010.00329.x (2011). [DOI] [PubMed] [Google Scholar]
59. Stamatakis A. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688–2690, doi: 10.1093/bioinformatics/btl446 (2006). [DOI] [PubMed] [Google Scholar]
60. Abascal F., Zardoya R. & Posada D. ProtTest: selection of best-fit models of protein evolution. Bioinformatics 21, 2104–2105, doi: 10.1093/bioinformatics/bti263 (2005). [DOI] [PubMed] [Google Scholar]
61. Silvestro D. & Michalak I. raxmlGUI: a graphical front-end for RAxML. Org. Divers. Evol. 12, 335–337, doi: 10.1007/s13127-011-0056-0 (2012). [DOI] [Google Scholar]
62. Latysheva N., Junker V. L., Palmer W. J., Codd G. A. & Barker D. The evolution of nitrogen fixation in cyanobacteria. Bioinformatics 28, 603–606, doi: 10.1093/bioinformatics/bts008 (2012). [DOI] [PubMed] [Google Scholar]
63. Brocks J. J., Buick R., Summons R. E. & Logan G. A. A reconstruction of Archean biological diversity based on molecular fossils from the 2.78 to 2.45 billion-year-old Mount Bruce Supergroup, Hamersley Basin, Western Australia. Geochim. Cosmochim. Acta 67, 4321–4335 (2003). [Google Scholar]
64. Crowe S. A. et al. Atmospheric oxygenation three billion years ago. Nature 501, 535-+, doi: 10.1038/nature12426 (2013). [DOI] [PubMed] [Google Scholar]
65. Rasmussen B., Fletcher I. R., Brocks J. J. & Kilburn M. R. Reassessing the first appearance of eukaryotes and cyanobacteria. Nature 455, 1101–U1109, doi: 10.1038/nature07381 (2008). [DOI] [PubMed] [Google Scholar]
66. van Dok W. & Hart T. B. Akinete Differentiation in Anabaena circinalis (Cyanophyta). J. Phycol. 32, 557–565 (1996). [Google Scholar]
67. Tomitani A., Knoll A. H., Cavanaugh C. M. & Ohno T. The evolutionary diversification of cyanobacteria: Molecular-phylogenetic and paleontological perspectives. P. Natl. Acad. Sci. USA 103, 5442–5447, doi: 10.1073/pnas.0600999103 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
68. Zhang Y. & Golubic S. Endolithic microfossils (cyanophyta) from early Proterozoic stromatolites, Hebei, China. Acta Micropaleontol. Sin. 4, 1–3 (1987). [Google Scholar]
69. Sims P. A., Mann D. G. & Medlin L. K. Evolution of the diatoms: insights from fossil, biological and molecular data. Phycologia 45, 361–402, doi: 10.2216/05-22.1 (2006). [DOI] [Google Scholar]
70. Huelsenbeck J. P., Nielsen R. & Bollback J. P. Stochastic mapping of morphological characters. Syst. Biol. 52, 131–158, doi: 10.1080/10635150390192780 (2003). [DOI] [PubMed] [Google Scholar]
71. Bollback J. P. SIMMAP: stochastic character mapping of discrete traits on phylogenies. BMC Bioinformatics 7, 88, doi: 10.1186/1471-2105-7-88 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
72. Castenholz R. W. et al. Phylum BX. Cyanobacteria. In: Bergey’s Manual® of Systematic Bacteriology, Springer New York. New York edn (2001). [Google Scholar]
73. Fralick P., Davis D. W. & Kissin S. A. The age of the Gunflint Formation, Ontario, Canada: single zircon U–Pb age determinations from reworked volcanic ash. Can. J. Earth Sci. 39, 1085–1091, doi: 10.1139/e02-028 (2011). [DOI] [Google Scholar]

Supplementary Materials