Heterotrophic flagellates are one of the cornerstones of aquatic ecosystems, as they are the prime bacterial consumers in the microbial food web and they are among the most important nutrient remineralizers and intermediators to higher levels of the trophic cascade [1
]. These organisms inhabit a variety of ecosystems, including low-level oxygen or anoxic bottom waters [3
], deep-sea sediments [6
], and hypersaline environments [7
]. Free-living heterotrophic and mixotrophic stramenopiles represent an important part of pelagic food webs, due to their capability to consume bacteria or dissolved organic material. Moreover, in some habitats, they represent about one third of all heterotrophic flagellates and thus, majorly contribute to a big part of nutrient recycling in the environment [8
Stramenopiles are a highly variable group ranging from heterotrophic nanoprotists to giant phototrophic multicellular kelps [10
]. Stramenopila can be divided into obligate heterotrophs, the apparently plastid-less labyrinthulomycetes, oomycetes, opalinids, and bicosoecids [10
]; a plastid-containing group, mostly phototrophic Ochrophyta [13
], that also includes known non-photosynthetic species such as apochlorotic nonphagotrophic diatoms from the genera Nitzschia
] and Pseudonitzschia
]; and phagotrophic chrysophytes, e.g., Mallomonas annulate
], to name a few. Interestingly, the ochrophytes also retain their plastids when they lose their autotrophic lifestyle [19
], while convincing evidence for the presence of a plastid has not been reported for any of the other groups of stramenopiles, which are all non-photosynthetic. Their plastids are complex in origin, and are derived from a rhodophyte ancestor [20
]. There is quite some debate on the relationship between the different groups of algae in which complex plastids are found (cryptophytes, haptophytes, ochrophytes, and most of the plastid-containing alveolates), as well as on the number and sequence of plastid gains, and losses or transfers between these groups [22
]. Comparisons of the phylogenetic signal found in mitochondrial (mt), plastid, and nuclear genomes, as well as genome/transcriptome-wide targeting predictions, provide insights into this currently unresolved convolute [23
]. However, these analyses need to be based on representative sets of organisms, covering the true diversity of these phyla. Phylogenomic analyses support the monophyly of the ochrophytes, while the non-photosynthetic stramenopiles consist of Oomycota (which. together with Ochrophyta, form the Gyrista), and Bigyra, a monophyletic group containing all other known heterotrophic stramenopiles [26
]. While there are considerable genome sequence data and knowledge on the economically significant Oomycota [27
], the other heterotrophic/bacterivorous stramenopiles, e.g., Platysulcus
and representatives of the MAST-3 clade, are underrepresented in sequence databases and in terms of knowledge about their ecology and physiology, in contrast to their global abundance in aquatic ecosystems [28
]. Furthermore, Oomycota, due to their mainly saprophytic and parasitic life strategies, and due to their phylogenetic position as part of the Gyrista, should not be considered as the “true” representatives of heterotrophic stramenopiles, especially because many groups in the Bigyra, like species from Platysulcus
], predatoric species of Developayella
], and the MAST-4 [32
] clade, seem to be exclusively composed of bacteriovorous lineages.
Bicosoecids are heterotrophic flagellates that represent a basal lineage of heterokonts [5
]. Although this group has been assigned to unicellular photosynthetic stramenopiles due to their ultrastructure [33
], analyses of molecular data have proven that Bicosoecida (clade Bikosea) are closer to stem stramenopiles [35
]. Moreover, Bikosea were assigned to the group Bigyra with Blastocystis
and Placidozoa as the closely related taxa [26
]. The main ultrastructural character distinguishing bicosoecids from other stramenopiles is the architecture of the feeding basket/flagellar apparatus, which was first thoroughly examined in Cafeteria roenbergensis
and has been established as an identifying feature since that time [30
]. Nevertheless, the feeding basket/flagellar apparatus is an ancestral motif, conserved in two different forms in bicosoecids and chromulinalean chrysophytes, but lost in other stramenopiles [35
], which lack the flagellar apparatus consisting of microtubular structures. The microtubular root (R3) has not been observed in any other stramenopiles in the sense of forming a prominent loop around the cell and hence it represents the main ultrastructural feature of Bicosoecida.
Due to the limited availability of sequence data from known species, molecular phylogenies are mostly restricted to a small ribosomal subunit of the nuclear RNAs (18S). The 18S rDNA sequences in combination with the feeding basked morphology are suitable as a basis for initial species recognition (see [5
]). To increase the sampling of heterotrophic stramenopiles, we isolated and cultivated a bicosoecid-like nanoflagellate from rock surface material collected in Norway. To determine its phylogenetic position more precisely, we studied the ultrastructure of this isolate and sequenced its mt genome.
We conducted a thorough morphological, ultrastructural, and molecular examination of the novel marine heterotrophic nanoflagellate C. marina.
It feeds on bacteria phagotrophically and belongs to the bicosoecids. Morphological and ultrastructural features typically described for bicosoecid cells—two flagella, the flagellar apparatus, microtubular root R3, and the tubular cristae in mitochondria [7
]—were observed in C. marina
. The bicosoecid lives and glides through the mucilage produced by a pelagophyte alga, with which it is in close association. A characteristic gliding/tumbling movement, with the anterior flagellum free and sweeping, while the posterior one was used as an anchor attached to the surface, was observed during cultivation. This type of movement has also been described for other bacteria-ingesting bicosoecids [62
]. According to our observations, C. marina
is a permanent feeder. Although the cytostome has not been seen in EM pictures, we assume that it is permanently present; firstly, due to the presence of the bacteria in the food vacuole (Figure 2
c black arrow), and secondly, due to the similarity of the life style (bacteriovory in marine sediments?) of Caecitellus parvulus
] and Rictus lutensis
], which are two related bicosoecids with cytostomes. This supposition is corroborated by the presence of R3 and ‘x’, which are structures typically associated with support of the feeding lip and the starting point of the cytostome [37
]. Available studies of bicosoecids have shown that they have four microtubular roots [37
]. For the description of ultrastructural characters, we used previously established terminology [37
]. The flagellar apparatus of C. marina
is similar to C. parvulus
and R. lutensis
]; however, some tubular features are different/missing. The same ultrastructural characters of the flagellar apparatus are found in C. marina
, C. parvulus
, R. lutensis
, and C. roenbergensis
; however, the number of microtubules for roots differs from the above named species. The most visible difference is in the R3 root: C. parvulus
has ~8–24 microtubules; R. lutensis
has ~4–50 microtubules; and C. roenbergensis
and C. marina
have 8 microtubules. Some of the species also exhibit other unique features, e.g., R. lutensis
has two additional microtubular structures ‘x’ and ‘S’, both of which are single-microtubule structures, whilst C. roenbergensis
is so far the only species with secondary cytoskeletal microtubules on the R4 root (see Table 1
for details). The number of microtubules per root is thus unique in C. marina
A peculiar major ultrastructural phenomenon of C. marina
is the tight connection of the mitochondria and nucleus in young cultures (Figure 2
). Although the clustering of mitochondria near the nucleus was documented for young rat heart cells [67
], endothelial cells of pulmonary arteries in mammalian cells [68
], mitochondria located in neurons [69
], and in thymocytes and leukemia cell lines [70
], a full conjunction of these two compartments, as seen in Figure 2
, has not yet been described. This organellar conjunction might be parallel to the known tethering and synergy between the mitochondria and endoplasmic reticulum (ER) [71
]. Their connection is involved in the regulation of lipid synthesis, Ca2+
signaling, and transport from ER to mitochondria [71
]. The bonding between the two compartments directly affects the mitochondria biogenesis [74
]. According to previous studies and results, we formed five hypotheses to explain this peculiarity: (i) The proximity of organelles might reflect intense ATP/ADP exchange between the mitochondria and the nucleus, via energy-dependent direct transport. The junction could serve as an energy "bridge" to cover the high energy demand of the nucleus during rapid cell growth and division [67
]. A similar provision of energy has been described for multiple cell lines/types, e.g., cardiac cells [67
] or endothelial cells of pulmonary arteries [68
]; (ii) the connection might facilitate the direct transport of tRNAs that are not encoded by the mt genome, but are required for mitochondrial translation. This would be similar to the situation in Trypanosoma brucei
, in which all mitochondrial tRNAs are encoded in the nucleus and are actively transported into the mitochondrion [76
]. In the case of C. marina
, alanine, glycine, and threonine (the amino acids carried by the exclusively nuclear-encoded tRNAs tRNAAla
, and tRNAThr
) are quite abundant (4.8%, 6.4%, and 5.9% of the predicted proteome, respectively) in the mitochondria-encoded proteins, hence the corresponding tRNAs must be imported into the mitochondrion; (iii) we should also take into consideration the hypothetical direct export of mRNA from the nucleus to the mitochondria as a template for the mitochondrial protein translation. Transport of mRNAs encoding mitochondrial proteins to the mitochondrial membrane and their cytosolic transcription is, in principle, known, but some mechanisms and associated processes are still not fully understood and explained [78
]. Regarding the mRNA transport hypothesis, it is interesting to know that the sharing of mRNA molecules between cells has been described for placental cells under oxidative stress in the syncytial tissue; mRNAs transcribed by active cells can then diffuse freely throughout the syncytioplasm into transcriptionally-inactive cells of the syncytium [80
]. However, currently, there is no known molecular machinery that can specifically recognize and transport mRNAs; (iv) the observed junction could also represent a physical connection of mitochondria and the nucleus during cell division to mechanically ensure equal segregation into the daughter cells. These cellular-level changes in organelle organization might also be linked to the variety of mitochondrial shapes, or be a special case of "kiss-and-run" mitochondrial motility dynamics, similar to the ones described by [81
]; (v) last but not least, the observed junctions might simply be a consequence of the small cell size and efficient intracellular organization. Because of the small cell size, the ER is extremely reduced in C. marina
. Since the nuclear envelope is continuous with the ER.; the observed junctions might also represent mitochondria-associated membranes. These are contact sites between ER and mitochondrial membranes [82
], that in the case of C. marina
, might reside directly at the outer side of the nuclear envelope.
In addition to the nucleus–mitochondria connections, we observed changes in the shape of the C. marina
Golgi apparatus during the cell cycle (Figure 3
). Interestingly, different forms and arrangements of Golgi cisternae seem to correlate with different cell cycle stages. A similar phenomenon was predicted for mammalian cells [58
] and later observed in vivo [57
]. While the flattening and curvature of the human Golgi has been associated with membrane curvature generators and changes in sphingomyelin metabolism [57
], in the case of the described nanoflagellate, the mechanism remains elusive.
The performed phylogenetic and morphological analyses with the comparison to other studies showed that C. parvulus
and the representatives of the family Cafeteriidae (Cafeteria
sp. and Cafeteria mylnikovii
) are the closest relatives to the novel bicosoecid species. Although the morphological features showed more similarities to C. roenbergensis,
namely the microtubule number of individual roots, shown in Table 1
, both types of molecular data (mt genome and 18S) share more similarities with C. parvulus
. The results of our phylogenetic analyses showed that C. marina
is not only a newly discovered species, but its divergence also qualifies it for placement in a new genus. Moreover, the already mentioned paraphyly of the genus Cafeteria
] and [26
] was confirmed by our analyses. However, this assumption is only based on the position of a single sequence of Pseudobodo tremulans
] and of Cafeteria
sp. in [26
], and further investigation is thus needed to confirm this claim.
Besides the C. roenbergensis mt genome, there is no other larger-scale bicosoecid data available for comparison. Our analyses showed a similar composition of the C. marina mt genome to the C. roenbergensis mt genome; however, with a different gene arrangement. The features typical for all stramenopiles, the lack of tRNAThr and the presence of suppressor tRNA consanguineous to the UGA stop codon, were also found in the C. roenbergensis mt genome, but the lack of tRNAAla and tRNAGly had been unique to C. marina up to now.
To summarize our findings, phylogenetic and morphological data concordantly place Cafileria marina into the family Bicosoecidae (Stein, 1878), with enough diverging evidence from known taxa to propose C. marina as a new species and genus of this family.
4.1. Taxonomic Summary
Cafileria marina is a heterotrophic stramenopile feeding on bacteria and is described under the International Code of Zoological Nomenclature.
4.2. Cafileria n. gen.
Flattened pyriform cells with two naked flagella, displaying tumbling movement with one flagellum attached to the substrate and the other flagellum moving. They have no lorica, cell wall, or other surface structures, and feed on bacteria by phagotrophy.
4.3. Type species
Cafileria marina n. sp.
4.4. Type locality
A three-meter depth off the shore in Gaustad, Norway (Kvernesfjorden fjord, 62°59′07.9″N 7°19′17.4″E), was used in this study.
4.5. Cafileria marina n. sp.
Cells 3–4 µm wide and 5–6 µm long, with anterior and posterior flagella of equal length that were 1.5–2 times length of the body, were employed. The cells and flagella had a smooth surface. The flagellar apparatus consisted of four roots (R1–R4) that did not have any secondary cytoskeletal microtubules.
Cafileria marina is named after the “kafilerie”, the Czech name of the place where the biomass of animal origin is dismantled and conditioned for the production of lipids, glue, and fertilizers. Therefore, we have seen a parallel with the feeding habits of the newly discovered nanoflagellate, which feasts on bacteria that recycle organic materials found in the biofilm mucillage that makes up their habitat. The species name marina stands for the marine origin of the species.
The designated hapantotype has been deposited as TEM and SEM specimens, and cryo-mixed cultures of nanoflagellate C. marina and bacteria S. litoris in the slide collection at the Biological Centre of the Czech Academy of Sciences, Institute of Parasitology, Ceske Budejovice, Czech Republic, under the accession number IP CAS Pro 59.
Eukaryota; Stramenopila; Bicosoecida.