Overlooked and widespread pennate diatom-diazotroph symbioses in the sea

Persistent nitrogen depletion in sunlit open ocean waters provides a favorable ecological niche for nitrogen-fixing (diazotrophic) cyanobacteria, some of which associate symbiotically with eukaryotic algae. All known marine examples of these symbioses have involved either centric diatom or haptophyte hosts. We report here the discovery and characterization of two distinct marine pennate diatom-diazotroph symbioses, which until now had only been observed in freshwater environments. Rhopalodiaceae diatoms Epithemia pelagica sp. nov. and Epithemia catenata sp. nov. were isolated repeatedly from the subtropical North Pacific Ocean, and analysis of sequence libraries reveals a global distribution. These symbioses likely escaped attention because the endosymbionts lack fluorescent photopigments, have nifH gene sequences similar to those of free-living unicellular cyanobacteria, and are lost in nitrogen-replete medium. Marine Rhopalodiaceae-diazotroph symbioses are a previously overlooked but widespread source of bioavailable nitrogen in marine habitats and provide new, easily cultured model organisms for the study of organelle evolution.

The internal view shows delicate costae running from margin to margin, continuing across the raphe where they are thicker in the form of transverse ridges, more pronounced near the center and poles of the valve (Supplementary Figs. 1m, n; 2b, d, f). One or two central costae may be shortened, not reaching the ventral side ( Supplementary Fig. 1m, n). The proximal raphe not visible ( Supplementary Fig. 2b). The distal raphe ends at the valve poles beyond the last thickened costa (Supplementary Fig. 2f). Two rows of slit openings visible between costae, forming a single uniseriate stria ( Supplementary Fig. 2d-

Description
Observation by LM of cells with protoplast: Cells joined together in chains, straight or curved, containing up to 60-70 cells (Fig. 1g). Cells connected by their valves in chains visible in girdle view (Fig. 1e). Girdles wider than valves, appear slightly rhomboidal (  Bacteria were attached to organic coating of uncleaned frustules ( Supplementary Fig. 4 a-g). 6 Observation by TEM of frustules: Hydrogen peroxide-cleaned valves have been observed with TEM. Valves elliptical with distinct axial slightly sigmoid keel ( Supplementary Fig. 5a, c).
Raphe with long fibulae ( Supplementary Fig. 5b, d). Raphe branches separated by a central node ( Supplementary Fig. 5a-c), and distal raphe ends with branching structure (Supplementary Fig.   5e, f). Microfibrils with unknown chemical composition were the only visible structural components of the valve walls ( Supplementary Fig. 5g), but it is not clear if they were associated with the surface of the silica (and secreted from the cell) or incorporated within it ( Supplementary Fig. 5b, c, e, g).
Reference culture: This novel species has been described from culture UHM3210. Strain show the overlap between both cingula, stars indicate the branches of the keeled raphe. e, 11 connection between two frustules by interlocking raphe keels of each valve (arrow). g, single frustule in valve view-note the axial position of the raphe keel and the central node (arrow). h, i, interval valve views (arrow). j, disintegrated frustule showing raphe central node (arrow) and one of the raphe branches with fibulae and branched distal end (star). k, element of fibulae.
Images a-g are from cells with protoplast, and bacteria are visible on the surface of diatom cells.

13
Supplementary Fig. 6. Light micrographs of live E. pelagica UHM3200 cells in valve view, imaged using a Nikon Eclipse 90i microscope at 60x magnification. A subset of the cells used for size measurements is included to illustrate possible morphological variability. Images were taken from two independent, exponentially growing cultures.

μm Epithemia pelagica UHM3200
14 Supplementary Fig. 7. Light micrographs of live E. pelagica UHM3201 cells in valve view, imaged using a Nikon Eclipse 90i microscope at 60x magnification. A subset of the cells used for size measurements is included to illustrate possible morphological variability. Images were taken from two independent, exponentially growing cultures.

μm
Epithemia pelagica UHM3201 Supplementary Fig. 8. Light micrographs of live E. pelagica UHM3202 cells in valve view, imaged using a Nikon Eclipse 90i microscope at 60x magnification. A subset of the cells used for size measurements is included to illustrate possible morphological variability. Images were taken from two independent, exponentially growing cultures.

μm
Epithemia pelagica UHM3202 Supplementary Fig. 9. Light micrographs of live E. pelagica UHM3203 cells in valve view, imaged using a Nikon Eclipse 90i microscope at 60x magnification. A subset of the cells used for size measurements is included to illustrate possible morphological variability. Images were taken from two independent, exponentially growing cultures.

μm
Epithemia pelagica UHM3203 Supplementary Fig. 10. Light micrographs of live E. pelagica UHM3204 cells in valve view, imaged using a Nikon Eclipse 90i microscope at 60x magnification. A subset of the cells used for size measurements is included to illustrate possible morphological variability. Images were taken from two independent, exponentially growing cultures.

μm
Epithemia pelagica UHM3204 Supplementary Fig. 11. Light micrographs of live E. catenata UHM3210 cells in valve view, imaged using a Nikon Eclipse 90i microscope at 60x magnification. A subset of the cells used for size measurements is included to illustrate possible morphological variability. Images were taken from two independent, exponentially growing cultures.

μm
Epithemia catenata UHM3210 Supplementary Fig. 12. Light micrographs of live E. catenata UHM3211 cells in valve view, imaged using a Nikon Eclipse 90i microscope at 60x magnification. A subset of the cells used for size measurements is included to illustrate possible morphological variability. Images were taken from two independent, exponentially growing cultures.   a-d, light micrographs of E. pelagica (a, b) and E. catenata (c, d), grown in either low-nitrogen medium (a, c) or nitrogen- replete K medium (b, d).
Cell samples were osmotically shocked with ultrapure water to improve visualization of the endosymbionts. All scale bars are 10 μm and black arrows indicate the presence of endosymbionts. The long-term growth experiments represented here, which investigate the loss of endosymbionts from cells grown in nitrogen-replete medium, were performed once for each E. pelagica and E. catenata. e, gel electrophoretic analysis of PCR amplifications performed on E. pelagica DNA extracts (lanes 2,4,8,10,14,16,20,22) and E. catenata DNA extracts (lanes 3,5,9,11,15,17,21,23) from cultures grown in either low-nitrogen medium (lanes 2, 3,8,9,14,15,20,21) or nitrogen-replete medium (lanes 4, 5, 10, 11, 16, 17, 22, 23). Diatom host marker genes 18S rRNA (lanes 2-6; 1,771 bp amplicon) and psbC (lanes 8-12; 1,252 bp amplicon) were successfully amplified from all culture DNA extracts, while the endosymbiont marker genes 16S rRNA (lanes 14-18; 458 bp amplicon) and nifH (lanes 20-24; 802 bp amplicon) were only successfully amplified from cultures grown in low-nitrogen medium. The psbC primer set is diatom-specific, while the 16S rRNA and nifH primer sets specifically target relatives of unicellular cyanobacterial diazotrophs (including Epithemia SBs). Note that in the case of nifH, the absence of appropriate template DNA led to increased non-specific amplification (lanes 22 and 23) but no amplification of the target genes. DNA ladders are provided in lanes 1, 7, 13, and 19, and the results of no template control (NTC) reactions for each gene are shown in lanes 6, 12, 18, and 24. For each gene, PCR reactions were prepared with 10 ng of template DNA (except for NTCs) using the same master mixes, and reactions were amplified for 35 cycles. Uncropped gel images are provided in the Source Data file.  (asterisks in the table accession numbers represent wildcards). Sequences previously identified as either "UCYN-C", "Cyanothece-like", or "unicellular cyanobacteria" are marked by filled circles; the remaining environmental sequences are either unpublished or were not identified in the associated study. Label colors correspond to the geographic origin of the samples, as highlighted on the map (note: the size of the highlighted boxes is arbitrary). The coordinates for these samples are plotted, except for KX064723 and KX064724 (from the South China Sea) and MH144515 and MH144458 (from the western North Pacific Ocean), where this information was unavailable. Bootstrap support values (% of 1000 replicates) are provided for major branches, and the scale bar is in units of nucleotide substitutions per site. Accession numbers for all sequences are provided in the Source Data file. Supplementary Fig. 24. Alignment of EcSB and EpSB nifH sequences with previously published primers and probes targeting the UCYN-C clade 7,15,16 . Bases in red represent disagreements with both the consensus sequence and primer/probe sequences. The primer and probe sequences displayed here have been reoriented for the alignment and thus may represent reverse complements of published sequences.  E. pelagica UHM3201 Suppl. Fig. 7 Suppl. Fig. 1a-h E. pelagica UHM3202 Suppl. Fig. 8 E. pelagica UHM3203 Suppl. Fig. 9 E. pelagica UHM3204 Suppl.