Unicellular Cyanobacterium Symbiotic with a Single-Celled Eukaryotic Alga

See allHide authors and affiliations

Science  21 Sep 2012:
Vol. 337, Issue 6101, pp. 1546-1550
DOI: 10.1126/science.1222700

Fixing on a Marine Partnership

Nitrogen fixation by microorganisms determines the productivity of the biosphere. Although plants photosynthesize by virtue of the ancient incorporation of cyanobacteria to form chloroplasts, no equivalent endosymbiotic event has occurred for nitrogen fixation. Nevertheless, in terrestrial environments, nitrogen-fixing symbioses between bacteria and plants, for example, are common. Thompson et al. (p. 1546) noticed that the ubiquitous marine cyanobacterium UCYN-A has an unusually streamlined genome lacking components of the photosynthetic machinery and central carbon metabolism—all suggestive of being an obligate symbiont. By using gentle filtration methods for raw seawater, a tiny eukaryote partner for UCYN-A of less than 3-µm in diameter was discovered. The bacterium sits on the cell wall of this calcifying picoeukaryote, donating fixed nitrogen and receiving fixed carbon in return.


Symbioses between nitrogen (N)2–fixing prokaryotes and photosynthetic eukaryotes are important for nitrogen acquisition in N-limited environments. Recently, a widely distributed planktonic uncultured nitrogen-fixing cyanobacterium (UCYN-A) was found to have unprecedented genome reduction, including the lack of oxygen-evolving photosystem II and the tricarboxylic acid cycle, which suggested partnership in a symbiosis. We showed that UCYN-A has a symbiotic association with a unicellular prymnesiophyte, closely related to calcifying taxa present in the fossil record. The partnership is mutualistic, because the prymnesiophyte receives fixed N in exchange for transferring fixed carbon to UCYN-A. This unusual partnership between a cyanobacterium and a unicellular alga is a model for symbiosis and is analogous to plastid and organismal evolution, and if calcifying, may have important implications for past and present oceanic N2 fixation.

Nitrogen (N) is a primary nutrient whose availability constrains the productivity of the biosphere (1). Some bacteria and archaea can fix N2 into biologically available ammonium and are important in the N cycle of terrestrial ecosystems and the global ocean. Although photosynthetic carbon (C) fixation evolved in eukaryotes through endosymbiosis of cyanobacteria that resulted in the chloroplast, no N2-fixing plastids or N2-fixing eukaryotes are known. Nonetheless, N2-fixing symbioses are common in terrestrial environments between bacteria or cyanobacteria and multicellular plants. In the oceans, there are microscopic observations of probable symbioses between N2-fixing cyanobacteria and single-celled eukaryotic algae (2), although the nature of the interactions, if any, are unclear, except between the heterocystous N2-fixing cyanobacteria and their associated diatoms (3).

Recently, a geographically widespread uncultivated diazotrophic cyanobacterium (UCYN-A) (4) was found to have an unusual degree of genomic streamlining suggestive of obligate symbiosis. The streamlined genome of UCYN-A (1.44 million base pairs) lacks photosystem II (PS II: the oxygen-evolving component of the photosynthetic apparatus), RuBisCo (ribulose-1,5-bisphosphate carboxylase-oxygenase that fixes CO2), and the tricarboxylic acid cycle (TCA), features that usually define cyanobacteria (5, 6). UCYN-A requires organic carbon for energy (although it may obtain some energy through cyclic photophosphorylation around PS I) and biosynthesis, as well as a number of specific amino acids and nucleotides (5, 6). We propose the name Candidatus Atelocyanobacterium thalassa for UCYN-A. Dissolved organic carbon concentrations in the ocean are typically low, particularly for labile compounds such as glucose that UCYN-A would require (it has a complete glycolysis pathway). Because UCYN-A has a complete suite of nitrogenase genes and related genes required for nitrogen fixation, it was hypothesized that UCYN-A provides fixed nitrogen in exchange for fixed carbon from a symbiotic partner (5, 7).

We tested different possible partner phytoplankton populations in seawater samples from the North Pacific Ocean (Fig. 1, fig. S1, and table S1) for the presence of symbiotic UCYN-A by screening flow-cytometrically sorted cells with a UCYN-A–specific quantitative PCR (qPCR) assay for the nitrogenase gene (nifH) (8). The UCYN-A genome had been obtained with this approach (6), but the seawater was first concentrated by vacuum filtration before sorting, which dislodged the UCYN-A from their associated cells (fig. S1). Here, we used a similar flow sorting procedure, but instead, raw seawater that had not been preserved or concentrated was immediately sorted and resulted in detection of most (63 to 94%) of the UCYN-A nifH genes in the sorted photosynthetic picoeukaryote population (PPE) (1- to 3-μm-diameter cells) rather than other pigmented and nonpigmented cells (sorted populations displayed in Fig. 1). The data unequivocally showed that UCYN-A is associated with photosynthetic picoeukaryotic cells. These results explain the reports of UCYN-A nifH in filter-fractionated samples from the California Coast (0.8- to 3-μm and 3- to 200-μm fractions) (9) and Station ALOHA (1- to 3-μm fraction) (5). We now only observe an enriched population of free UCYN-A cells (0.2 to 1 μm) after seawater concentration by vacuum filtration, freezing for storage purposes, and resuspension in sterile seawater, which apparently disrupts the fragile association (fig. S1). UCYN-A appears to be in a loose extracellular association (epiphytic) that is easily dislodged, which explains why there is some amplification of UCYN-A nifH from outside the region of the sorted photosynthetic picoeukaryote population (table S1). This delicate association is similar to microscopic observations of other probable marine plankton symbioses, including the mixed populations of unicellular Synechococcus- and Crocosphaera-like unicellular cyanobacteria housed in the girdle of Dinophysis (dinoflagellates) (10).

Fig. 1

Example flow cytogram of cell populations in unpreserved seawater that were targeted in this study. Red fluorescence is a measure of chlorophyll a concentration per cell. Forward scatter (FSC) is a proxy for cell size. Beads 3 μm in diameter (black) were used for reference. Cell populations indicated are photosynthetic picoeukaryotes (PPE, blue), Prochlorococcus (Pro., green.), and cells (gray) that are not PPE or Pro. Coloring of each population indicates the sort gates used. Most UCYN-A nifH gene copies (63 to 94%) were amplified from the PPE population (table S1). Flow cytograms of all samples used in this study are presented in fig. S1.

The marine picoeukaryotic population defined by flow cytometry is extremely diverse (1113). Therefore, to identify the specific cells associated with UCYN-A, we compared universal 18S ribosomal RNA (rRNA) and 16S rRNA gene clone libraries that were amplified from sorted samples of the entire picoeukaryote population to those from sorted single picoeukaryote cells (table S1). Single cells and the entire picoeukaryote population were initially screened for UCYN-A by nifH qPCR. As expected, the partial 18S rRNA gene sequences [~730 base pairs (bp)] derived from the entire picoeukaryote population sorts were diverse and included sequences from several classes of marine picoeukaryotes (Fig. 2 and table S3). 16S rRNA gene sequences amplified from the entire picoeukaryote population sorts were also diverse and confirmed the presence of UCYN-A in these populations (table S3). However, amplification of the 18S rRNA gene (using nested PCR) from single UCYN-A nifH-positive sorted picoeukaryote cells yielded exclusively prymnesiophyte sequences (Fig. 2 and table S4).

Fig. 2

Diversity and phylogeny of UCYN-A nifH-positive sorted picoeukaryote populations and single cells from 5-m and 79-m depths at Station ALOHA. (A) Maximum-likelihood tree (PhyML) of selected cultured prymnesiophyte 18S rRNA gene sequences with clade assignments and genus names per (15). “BIOSOPE T60.34” sequence (blue type and asterisk) is the best BLASTn hit of the UCYN-A partner sequence amplified from UCYN-A nifH-positive single photosynthetic picoeukaryotes. Node support greater than 75% is marked by black squares. (B) 18S rRNA gene diversity of the entire sorted picoeukaryote population that was the source of single cell sorts from March (5-m depth, cruise KM1110) and August (79-m depth, cruise HOT234) that were positive for UCYN-A nifH and the BIOSOPE T60.34 sequence.

The partial 18S rRNA gene sequences [or “partner” sequences from the Ek555/1269 primer amplicon, GenBank accession nos. JX291679 to JX291804 and JX291547 to JX291678], derived from 12 single UCYN-A–nifH–positive picoeukaryotic cells, were greater than 99.5% identical to each other and had best BLASTn hits (>99% identical) to a sequence derived from sorted picoplankton from the oligotrophic South Pacific Ocean (GenBank accession no. FJ537341, BIOSOPE T60.34, sample T60, Station STB11, –107.29°E, –27.77°S) (14) (Fig. 2). A metagenome from sorted photosynthetic picoeukaryotes of the same sample (BIOSOPE, sample T60) also contained numerous DNA sequence reads (mean length of 340 bp) that were almost identical (99 to 100%) to the UCYN-A genome. Samples from an adjacent station (STB7, samples T39 and T40) contained neither good matches to the UCYN-A genome nor the partner partial 18S rRNA gene sequence we identified (table S2 and fig. S2). Assembly of the UCYN-A reads from sample T60 covered 12.4% of the UCYN-A genome (table S2), providing additional evidence that UCYN-A is associated with the picoeukaryotes and that the UCYN-A partner is the same in the oligotrophic North and South Pacific Oceans.

We examined the phylogeny of the full-length 18S rRNA gene of the BIOSOPE environmental sequence (T60.34) (UCYN-A partner sequence best BLASTn hit) relative to the diverse prymnesiophyte class (11, 15, 16). The BIOSOPE T60.34 sequence clustered with the calcareous nanoplankton Braarudosphaera bigelowii (1720) and with Chrysochromulina parkeae (21, 22), which may contain calcified scales as well (23) (Fig. 2). B. bigelowii appears to represent several pseudo-cryptic species that are all easily differentiated from other calcareous phytoplankton by their distinct pentagonal plates (19). Production of calcified plates by the UCYN-A partner is an intriguing possibility, because calcareous phytoplankton are important in the vertical flux of carbon and nitrogen in the oceans. Sedimentary records from the late Cretaceous show fossils of Braarudosphaera species in open ocean sediments (24, 25), suggesting that the UCYN-A symbiosis could be ancient and could potentially be studied in a paleo-oceanographic context. However, prymnesiophytes (for example, Emiliania huxleyi) have complex life histories in which form (in particular calcification) and behavior change dramatically between haploid and diploid life stages [references in (26)]. Because the life-history stage of the partner cells in the natural populations of this study is unknown and sample processing could have dissolved or removed plates, whether or not the partner is calcifying could not be determined. Calcification of the partner cell could be an important facet in its symbiosis with UCYN-A, because it could provide a mechanism for stabilizing an extracellular association.

Prymnesiophytes are typically free-living and photosynthetic and therefore could provide organic C for an associated photoheterotroph like UCYN-A. To test this hypothesis, we applied a halogenated in situ hybridization nanometer-scale secondary ion mass spectrometry (HISH-SIMS) approach to natural phytoplankton populations from Station ALOHA (27) that were amended with 15N2 and 13C-bicarbonate (H13CO3) and incubated under in situ conditions. The photosynthetic picoeukaryote cells (diameter 1 to 3 μm) were subsequently sorted by flow cytometry, preserved, and processed for HISH-SIMS (27, 28). A highly specific oligonucleotide probe for the UCYN-A 16S rRNA gene (UCYN-732) (27) confirmed the presence of 278 UCYN-A cells among the sorted photosynthetic picoeukaryote population. Most (163, 59%) of the UCYN-A cells (diameter 0.31 to 0.92 μm) remained associated with a larger partner cell (diameter 0.99 to 1.76 μm) during sample processing. UCYN-A cells were mainly observed at one end of the partner cell in what appeared to be an indentation (Fig. 3A). Numerous UCYN-A cells (107, 38%) were dislodged from their eukaryotic partners during HISH-SIMS sample processing (the cells were initially attached, otherwise they would not have been sorted by flow cytometry) and observed without an associated cell. Some (3%) of the UCYN-A cells identified by HISH were present as pairs located at opposite ends of a single partner cell (Fig. 3A). In other planktonic symbioses involving cyanobacteria (e.g., diatom symbionts with filamentous heterocyst-forming cyanobacteria), the cyanobacteria migrate to polar ends of the larger partner cells before cell division of the host (29). Similarly, the partner cells with multiple associated UCYN-A cells may have been dividing, because the incubation period was longer than a typical cell division cycle (36 hours). Measurements of scanning electron micrographs (SEMs) of individual B. bigelowii cells from enrichments of coastal waters show three size classes, the smallest of which corresponds to cell diameters of about 5 μm, which is larger than the UCYN-A partner cell measured here (18, 19). If calcareous plates are present on the UCYN-A partner, loss of the plates during sample processing for HISH-SIMS may result in measurement of a smaller cell diameter than cells measured by SEM. Alternatively, the UCYN-A partner may represent an additional smaller-size class of B. bigelowii adapted to the oligotrophic ocean.

Ten partner cells and their associated UCYN-A cells were chosen for quantitative isotopic analysis with nanoSIMS. On average, 13C/12C enrichment was lower in the UCYN-A cells than in the partner cells (average 13C atom % of 1.8139 and 2.4602, respectively; Fig. 3, B and C, and table S6). Because the UCYN-A genome does not contain carbon-fixation pathways (5, 6), but photosynthetic picoeukaryotes do, we conclude that the 13C enrichment in UCYN-A was due to transfer of fixed C from the eukaryotic partner. Both partner and UCYN-A were strongly 15N-enriched (Fig. 3, C and D, and table S6). As only some bacteria and some archaea can fix N2, our nanoSIMS data provide direct evidence for active N2 fixation by the uncultivated UCYN-A. The average 15N/14N enrichment was higher in the partner cells than in the UCYN-A cells (average 15N atom % of 1.5308 and 1.2081, respectively), showing that extensive amounts (we estimate up to 95%) of fixed N were transferred from UCYN-A to the partner cell. In contrast, little of the C (1 to 17%) fixed by the partner cell was transferred to the UCYN-A, consistent with the lower C requirements of a small slow-growing heterotrophic symbiont. The large fraction of N transferred to the partner is consistent with results from known symbioses, such as between heterocystous cyanobacteria and marine planktonic diatoms (3).

The symbiosis reported here is unusual in that it is a partnership between a prymnesiophyte and a unicellular cyanobacterium. The results provide an explanation for how the metabolically streamlined UCYN-A survives in the oligotrophic ocean, despite the lack of the TCA cycle, PS II, and some biosynthetic pathways. Tracer experiments using 15N and 13C clearly show that the cyanobacterium provides fixed N to the eukaryotic partner and conversely that C fixation by the eukaryotic partner can provide C to UCYN-A. Thus, the association appears to be at least a mutualistic and facultative, if not an obligate, association. It is still not known whether the cyanobacterium is an endosymbiont or lives on the surface of the prymnesiophyte. However, the sensitivity to disruption and the HISH-SIMS imaging indicates that it is a loose cell-surface association. Many of the dislodged cells observed after HISH-SIMS were located near a picoeukaryotic cell. Because B. bigelowii, the closest known relative of the sequence amplified from our single-cell sorts, has calcareous plates that are easily dislodged, it is conceivable that UCYN-A is somehow associated with these plates. Sample handling and processing, in particular the HISH assay, could have dislodged or dissolved the calcium carbonate plates, releasing UCYN-A.

The association of UCYN-A with prymnesiophytes suggests that N fixed by UCYN-A may enter the microbial loop through this group of relatively abundant and globally relevant primary producers and mixotrophs (11, 16, 30, 31) and has other important implications. The UCYN-A partner may be calcifying, which has implications for the contributions of N2-based new production to vertical carbon fluxes (32), the sensitivity of the UCYN-A partner to ocean acidification (33), and paleo-oceanographic microfossils in sediments. Equally interesting is the existence of a planktonic unicellular N2-fixing symbiosis and its implications for evolution and adaptation. The presence of these simple interactions between single-celled organisms are reminiscent of those earlier primary endosymbiotic events and underscores the enigmatic absence of N2-fixing plastids in evolution as N2 fixation is an energetically expensive oxygen-sensitive reaction, and nitrogenase is an ancient enzyme. Thus, the UCYN-A association is a symbiosis with a prymnesiophyte and provides an intriguing model for the evolution of N2 fixation, and the mutualistic interactions between planktonic microorganisms.

Fig. 3

Microscopy and elemental composition of two UCYN-A partner cells and their associated UCYN-A cells detected in samples from sorted picoeukaryotes analyzed by HISH-SIMS. (A) 19F/12C (HISH) labeling of UCYN-A. Inset displays labeling of the same UCYN-A cells by catalyzed reporter deposition–fluorescence in situ hybridization (green) and DAPI (4′,6-diamidino-2-phenylindole) staining of partner cell nucleus (blue). (B) The 13C/12C ratio image of UCYN-A and partner cell. (C) The 13C/12C and 15N/14N in 10 selected partner cells and their associated UCYN-A cells (table S6). (D) The 15N/14N image ratio of UCYN-A and partner cell. The white lines define regions of interest that were used for calculating 13C/12C and 15N/14N ratios. UCYN-A cells are indicated by white arrows in (B) and (D). Scale bar, 3 μm.

Supplementary Materials

Materials and Methods

Figs. S1 and S2

Tables S1 to S6

References (3446)

  • Present address: Department of Isotope Biogeochemistry, UFZ–Helmholtz Centre for Environmental Research, Leipzig 04318, Germany.

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: D. Bottjer and M. Hogan provided advice for 15N2 additions. Water samples were collected with the help of S. Curless, M. Church, S. Wilson, S. Tozzi, and the captain and crew of the research vessel Kilo Moana. On-board flow cytometry was made possible by K. Doggett and D. Karl. Funding was provided by the Gordon and Betty Moore Foundation (J.P.Z.) and the NSF Center for Microbial Oceanography: Research and Education (C-MORE). The Max Planck Society sponsored the HISH-SIMS analysis. We thank G. Lavik (Max Planck Institute, Bremen) for advice and suggestions for data analysis. J. Waterbury provided the scientific name for UCYN-A. D.V. was supported by PHYTOMETAGENE (JST-CNRS), METAPICO (Genoscope), and Micro B3 (funded by the European Union, contract 287589). BIOSOPE metagenome sequencing was performed at Genoscope (French National Sequencing Center) by J. Poulain. We thank H. Claustre, A. Sciandra, D. Marie, and all other BIOSOPE cruise participants. GenBank accession nos.: JX291679 to JX291804 and JX291547 to JX291678 (see table S4 for details).
View Abstract

Navigate This Article