In popular terms, the marine food chain goes like this: "phytoplankton feed the zooplankton that feed the small fish (larvae) and crustaceans that feed the larger fish that feed the even bigger fish that feed us". Thus we have us, Homo sapiens, together with the great white shark, Carcharodon carcharias, and the orca, Orcinus orca, as the apex predators, the species with the highest trophic levels. With a low trophic level and at the other end of the size scale, "among the plankton, the food web is so tangled and complex even scientists don't know who eats whom" (Source, starting at 5:12 ). Wrong! Here thrives, as we now know, the tiniest imaginable unicellular eukaryotic predator, Braarudosphaera bigelowii (frontpage, Figure 1). In science lingo, one could call it the 'antapex predator'*) if that wouldn't sound downright bonkers. But see yourself...
The ball‑shaped cells of Braarudosphaera have an estimated diameter of 1.3 ±0.22 μm and a cell volume of 2.8 ±0.8 μm3 (Figure 1) They are thus only marginally larger than the smallest known unicellular eukaryote, Ostreococcus tauri, with its diameter of ~0.8 µm (featured by Elio here in STC ). Kamennaya et al. found one or two undulipodia (the functional equivalent of the bacterial 'flagella' ) on some of their 185 sampled Braarudosphaera cells, and, infrequently, organic, non-calcified scales (they can't be cultivated in the lab so far ). The cells are densely packed with the nucleus, one mitochondrion, the ribosome-rich endoplasmatic reticulum, two chloroplasts, and an endosymbiont (more on this endosymbiont below ). So densely, in fact, that the endosymbiont hardly fits inside and was seen often as a bulge at the cell periphery (Figure 1, H+K). Also, the stuffed inside of the cells would hardly allow food ingestion by conventional phagocytosis, that is, swallowing bacteria whole through their mouth-like cytostome. Roughly 80% of the Braarudosphaera cells in their sample were found to be attached to not-much-smaller cells with diameters of 0.81 ±0.08 μm, and which turned out to be Prochlorococcus. In electron micrographs, the Prochlorococcus cells appeared as half-swallowed (Figure 1, H–K), but they were also seen missing (Figure 1, B+C) or in the process of de-/attachment (Figure 1, D, and on the frontpage), which can be an artifact from cell preparation for microscopy. A greater percentage of the Prochlorococcus cells had doughnut-like shapes rather than their usually rounded shapes, which suggested that Braarudosphaera literally sucks its prey before disposing of the carcass. And apparently, Braarudosphaera has a penchant for Prochlorococcus as it was not found attached to and feasting on Pelagibacter cells that were also present in their samples. But careful! The Prochlorococcus cells might just fit better and thus stick more strongly to the cytostomes, like a plug, than the more elongated and crescent-shaped Pelagibacter cells (Kamennaya et al. propose the term 'Pomacytosis' for this funny semi-extracellular phagocytosis; the term is derived from the Greek πώμα for 'plug', but I would prefer to not follow them here to keep the science terminology from spilling over ). Anyways, members of the phytoplankton, usually considered autotrophic 'primary producers' at the base of the marine food web, can also engage in a rapacious heterotrophic life style. One at least does, Braarudosphaera, but there may well be more.
Kamennaya et al. found 99% identity for the 18S rRNA sequence of their tiny predator from the Eastern subtropical North Atlantic Ocean with that of one calcifying Braarudosphaera bigelowii isolate from the Japanese coast, studied by Hagino et al., and one non-calcifying photosynthetic picoeukaryote (<3 µm diameter, as yet unnamed ) detected by Shi et al. in samples from the South East Pacific Ocean. They concluded that their predator is a strain of B. bigelowii, or a sister species (I mention the different sampling locations to emphasize that they're apparently ubiquitous in the oceans ). But there's a twist, and that has to do with 'calcifying'. The coccolithophores come in different 'flavors' (we featured Emiliania huxleyi here in STC ), and it is known for several of their species that they can be either haploid or diploid, depending on their life cycle, and can switch between life cycle stages, between ploïdies, in response to environmental cues. They probably also have sex, which no one has seen so far but wouldn't make things easier, as usual. Many coccolithophores produce the most ingenious calcareous scales as exoskeletons, but not all of them do, and not necessarily during all their life cycle stages, and not always scales of the same type in different stages. For example, Chrysochromulina parkeae is a coccolithophore of the Prymnesiophyceae family that produces spikes rather than 'pentalith' scales yet is so closely related to B. bigelowii (99.89% 18S rRNA identity) that it might well be that both represent different life cycle forms of one single species. There you have it: it's complicated (=messy, but not necessarily overly complex ). But coming back to B. bigelowii, Hagino et al. characterized several scaled isolates (Figure 2) that had, not the least due to their >1 µm thick armour, cell diameters of ~15 µm and thus entirely different dimensions than the tiny predators of Kamennaya et al.. And, to mention a nutritious detail, these scaled B. bigelowii (here's a stunning 'portrait gallery' ) were not found sucking on Prochlorococcus. For 'technical reasons', obviously.
B. bigelowii's 'spheroid body', or 'cyanobiont', is enthralling in its own right (Figure 3). Yes, it's a cyanobacterium, but no, it's not another chloroplast-type endosymbiont! Kamennaya et al. found by 16S rRNA‑typing that the endosymbiont is 100% identical to UCYN-A (unicellular diazotrophic cyanobacteria group A), later baptized 'Candidatus Atelocyanobacterium thalassa'; 'Candidatus' because it has not yet been grown in the lab but only enriched from sea water samples by flow-cytometry. With a diameter of 0.47 ±0.05 µm, the cell size of UCYN-A is similar to that of Prochlorococcus and Pelagibacter ubique (SAR11), which were, no surprise, also found in the enriched sample fraction. The 1.4 Mb-small genome of A. thalassa – also comparable in size to that of Prochlorococcus (1.66 Mb) or P. ubique (1.3 Mb) – contains cyanobacterial core genes, the complete nif gene cluster for nitrogen fixation, and the genes for photosystem I (PSI), but Zehr et al. did not find genes for carbon fixation, (=no carboxysomes, no RuBisCo ), or for the oxygen-producing photosystem II (PSII) and its associated pigments like, for example, phycoerythrin. The latter was a first for the cyanobacteria: non-phototrophic, that is, heterotrophic N2‑fixing cyanobacteria weren't known before, ten years ago. Also missing from the core genes of A. thalassa are the dnaA gene and an oriC-type replication origin, which are both key components for bacterial chromosome replication and easily identifiable in free-living cyanobacteria as, for example, in S. elongatus PCC 7942 or Synechococcus sp. PCC 7942, but are frequently missing or present as pseudogenes in symbiotic cyanos like Nostoc azollae 0708 (author's own unpublished observation ). Maybe such symbiotic cyanos have skipped their 'normal' daylight-triggered replication cycle to synchronize with the longer replication cycles of their algal or plant hosts? We don't know. But it's reasonable to assume that the A. thalassa cyanobiont of B. bigelowii supplies fixed nitrogen to the metabolism of the holobiont in exchange for 'carb'. Since also other algae of the phytoplankton 'symbiose' with N2‑fixing cyanos in the oceans, their joint contribution – and not only that of the free-living filamentous cyanobacterium Trichodesmium – to the global nitrogen cycle should be of similar importance as the oxygen production by other cyanobacteria – free-living ones or chloroplasts – to the global oxygen cycle. A facet of the wondrous cyanobacteria I wasn't aware of.
*) antapex: antonym (=counter term ) of 'apex', possibly from Greek πάγος, mountain peak.