Multifaceted Dinoflagellates and the Marine Model Prorocentrum cordatum

The collective performance of phytoplankton in the sunlit zone of the world’s oceans provides a major share of the global primary production, qualifying it as a relevant driver of the global biogeochemical cycles and the marine food web. Furthermore, phytoplankton members form harmful algal blooms (HABs, red tides), which emerge more frequently due to climate change and progressive eutrophication of marine systems, in particular in coastal ranges and shelf seas. HABs often lead to oxygen depletion and toxin accumulation, representing serious stressors to the marine ecosystem and causing significant losses to fisheries and aquaculture industries [1], exemplified by the loss of 4,700 t of salmon in 2016 following a HAB off the coast of southern Chile [2].

Dinoflagellates are ancient and abundant phytoplankton members, which (i) exhibit tremendous organismic and morphologic diversity, (ii) occur as free-living organisms, endosymbionts, or parasites of other protists and multicellular eukaryotes, and (iii) stand out for their unusually large genomes and various cellular peculiarities. Despite the generally acknowledged role of dinoflagellates for marine and freshwater ecosystems and their unique cellular features, the molecular basis of their physiology, adaptability, and cell biology are under-investigated and largely non-understood. Prorocentrum cordatum is a ubiquitously occurring photosynthetic HAB-former and represents a promising model for studying dinoflagellates on the systems biology level.

This review aims at summarizing recent insights into the cell biology and metabolism of P. cordatum framed by an overview of the current knowledge on dinoflagellates. Specific topics covered include: (i) biogeography, (ii) lifestyles, interactions, and algal blooms, (iii) evolution, phylogeny, and taxonomy, (iv) general morphology and subcellular structures, (v) genomes and genetic information processing, (vi) photosynthetic machinery and energy metabolism, and (vii) other model dinoflagellates.

Dinoflagellates are ubiquitous and plentiful members of marine and freshwater environments. It is estimated that ∼90% of the currently described dinoflagellate species inhabit marine waters, whereas only ∼10% has been recorded from freshwater environments [3, 4]. In the marine system, dinoflagellates are globally distributed and tremendously diverse with some oceanic regions particularly rich in them as, e.g., the Pacific-Indian Ocean compared to the tropical Atlantic [3]. Nevertheless, almost identical dinoflagellate communities were observed in similar climate zones of the northern and southern hemisphere [3]. A “belt of circumtropical species,” which inhabits the tropical waters around the equator, separates the southern and northern communities [3]. Examples for circumtropical species are members of the genera Ornithocercus and Histioneis [3]. Dinoflagellate species that inhabit different climate zones of the oceans were defined as ubiquitously occurring, such as members of the genera Prorocentrum [5], Tripos [6], or Alexandrium [7]. Species that are exclusively present in a single oceanic region are rare and mainly reported from polar ocean waters [3]. Knowledge about species, which were recorded from snow, sea ice, and sea water in polar regions are, e.g., freshwater dinoflagellates of the genera Gymnodinium [8‒10] as well as the marine Alexandrium catenella [11] and Polarella spp. [12, 13].

During the past decade, widespread application of 18S rRNA gene metabarcoding greatly expanded our knowledge on the global distribution patterns of dinoflagellates, as exemplified in the following: (i) Analysis of the TARA-Ocean database revealed a pronounced size-dependent structuring of the dinoflagellate communities, e.g., extreme diversity of Gymnodiniales in the pico- to nano-planktonic fractions [14]. (ii) Inspection of the metaPR2 v.1.0.3 database revealed high abundance of members of the family Suessiaceae [12]. (iii) Investigations on the distribution of micro-eukaryotic plankton across the complete west-east transect of the Pacific Ocean indicated Syndiniales (parasitic) to be ubiquitously distributed [15]. (iv) Profiling the epipelagial of the western Coral Sea showed a differential community profile with photosynthetic/heterotrophic dinoflagellates dominating in the upper and Syndiniales in the deeper layers [16].

P. cordatum is a ubiquitously occurring dinoflagellate and was observed in regions ranging from tropical and subtropical waters to temperate zones of the oceans [5]. Our present literature-based survey supports and broadens these previous findings, by outlining the number and distribution of P. cordatum in the world’s oceans (Fig. 1a; online suppl. Table S1; for all online suppl. material, see https://doi.org/10.1159/000540520). P. cordatum mainly occurs in estuarine and coastal waters with a few examples also recorded in the open ocean. Further, it inhabits all climate zones, except for Antarctic waters. P. cordatum was mostly observed or isolated from the Subarctic North Atlantic, the Subarctic North Pacific, and the Subarctic zone. The highest abundance of P. cordatum species was recorded for the Baltic Sea (Fig. 1b). Zoom-in graphics of the tropical temperate macrozone around North America and the subarctic North Pacific region off East China revealed the ample distribution of P. cordatum in coastal and estuarine waters (Fig. 1c, d).

Based on contemporary knowledge, the majority of marine dinoflagellates are free-living (1,555 species) [17], while the remainder has an endosymbiotic or parasitic lifestyle [3]. Free-living dinoflagellates are classified into planktonic or benthic species, and some of them were observed as abundant organisms in HABs [18, 19]. Examples for free-living dinoflagellates are members of the diverse genera Amphidinium, Tripos, or Prorocentrum [17]. Endosymbiotic, single-celled dinoflagellates (i.e., “zooxanthellae”) were predominantly observed in a mutualistic association with radiolarians and reef-building corals [20‒22]. This type of lifestyle is well studied for members of the family Symbiodiniaceae [20, 23, 24]. Only ∼5% of known dinoflagellates have maintained a parasitic lifestyle [3], which is known from, e.g., members of the genera Amoebophrya and Blastodinium [25, 26]. Parasitic dinoflagellates infect a broad variety of host organisms, including other protists, free-living dinoflagellates, invertebrates, and in some cases also vertebrates [27].

In total, around half of the currently described dinoflagellate species is described as photosynthetic, which includes autotrophic and mixotrophic species as well as heterotrophs [3, 28]. Furthermore, dinoflagellates are extremely diverse in terms of their metabolism [29, 30], pigment composition [31], ability to produce toxins [19, 32, 33], or capacity for bioluminescence [34].

In the sunlit, euphotic zones of the oceans (0–200 m water depth), dinoflagellates are part of the (micro-)phytoplankton along with diatoms, other microalgae (e.g., haptophytes), and cyanobacteria (Fig. 2a) [35]; cumulatively accounting for only ∼1% of the earth’s photosynthetic biomass. Despite this, they make an above-average contribution to the global net primary production (∼45%), due to their high turnover rates, seasonal stability, and the high photosynthetic performance per cell [35, 36]. By fixing atmospheric CO₂, phytoplankton contributes to governing the global climate [36, 37]. Furthermore, phytoplankton assimilates other inorganic nutrients into organic biomass (in particular nitrogen and phosphorus), releases dissolved organic matter (DOM), and serves as food source, e.g., for zooplankton and higher trophic levels (Fig. 2a) [38‒40]. Since dinoflagellates account for ∼50% of the total CO₂ fixed by phytoplankton, they are important key players in the carbon cycle of the marine ecosystem [41]. Likewise, mixotrophic dinoflagellates (e.g., Prorocentrum c.f. balticum [42] and P. cordatum [43‒45]) significantly contribute to the carbon flux in the oceans (Fig. 2b). During photosynthesis in daytime, P. cf. balticum produces a “carbon-rich” mucus trap (synonym: microsphere) in order to capture and consume prey at night [42]. Prey cells are either passively entangled in the mucus trap or actively drawn into it by an inward pulling force, which is generated by the rotating dinoflagellate cell [43]. Then, prey cells are fixed by the peduncle to the apical pole of the dinoflagellate cell, followed by phagocytosis into feed vacuoles [43]. Experiments with recently described Prorocentrum pervagatum [46] revealed an uptake rate of up to 10‒12 cells per day of the cryptophyte prey Teleaulax amphioxeia, yielding faster growth compared to the unfed control cultures [43]. The dinoflagellate releases the mucus trap when satiated, thereby initiating a downward flux of carbon as particulate organic carbon into deeper waters [42].

Furthermore, interactions of microalgae with, e.g., bacteria have been well studied. These associations are observed to be mutualistic [47‒49], pathogenic [50, 51], or both [52, 53]. In the case of P. cordatum, its associated microbiome was recently shown to possess a core community, which was rather stable irrespective of the geographical sampling site of the host [54]. Other studies showed that dinoflagellates interact with viruses and partially integrate viral genomes [55, 56]. Some dinoflagellates are able to produce toxins as protection against grazers [57]; others produce bioactive secondary metabolites, which are suggested to prevent the organisms from digestion by their host during endosymbiosis [58].

Algal blooms are natural phenomena, where algal populations rapidly increase (accumulate) in the surface layers of freshwater and marine water bodies. This accumulation can have positive effects on the ecosystem functions; however, the proliferation of, e.g., harmful, toxin-producing algae can become problematic. Furthermore, the collapse of algal blooms leads to massive releases of organic carbon, fueling heterotrophs to deplete O₂ and release CO₂.

Algal blooms in general occur naturally due to various stimulating environmental factors, which include higher nutrient availability (mineral and/or organic), rising water temperatures, and increasing light intensities [59‒62]. In particular, increasing organic and inorganic forms of nitrogen and phosphate are suggested to stimulate the formation of algal blooms [63, 64]. In the marine system, algal blooms predominantly arise in estuarine and coastal regions, where the aforementioned triggering factors are more prevalent than in the pelagic realm of open oceans [65‒68]. Algal blooms show seasonal variations as they arise several times a year with maxima in March and in August/September as observed during the last few years, e.g., in the Indian Ocean and in the northern coastal waters of the Persian Gulf and Oman Sea [68, 69].

The reason for an algal bloom to collapse is not monocausal, but rather linked to various factors such as nutrient limitation, viral infections, and/or grazing [70‒72]. During a collapse, massive amounts of organic material are released into the environment, which can have major impacts on the ecosystem. This organic material is subsequently degraded by heterotrophic microorganism, which at the same time consume large amounts of O₂, leading to the formation of anoxic zones [73‒75]. These so-called dead-zones have strongly negative effects on the viability of oxygen-respiring organisms, resulting in complete changes of the community structures [74]. Furthermore, release of organic carbon enriches DOM, which represents one of the largest carbon pools on Earth and comprises countless components with half-life times of several thousands of years [76].

In the course of climate change, next to temperature increase also nutrient availability (including increasing anthropogenic loading) in ocean waters is affected [77‒79]. This translates into shifts of surface ocean ecosystems as already inferred from ocean color trends [37]. Indeed, recent studies on the development and occurrence of coastal algal blooms have shown that they have expanded significantly and become more frequent during recent decades [66, 80]. Areas in the entire southern hemisphere are most affected by algal blooms, while these apparently decrease in tropical and subtropical regions of the northern hemisphere [66]. The study by Dai et al. [66] further documented a correlation between bloom trends, ocean circulation, and increasing sea surface temperatures [66]. In particular, the latter and rising light intensities in the context of climate change are suggested to promote the development of algal blooms [80, 81]. A recent study by Wolf et al. [82] on arctic phytoplankton found that the cooling in the aftermath of a heatwave likewise impacts the overall phytoplankton productivity. Further observations of increasing HAB formation were made for different oceanic regions including the North Atlantic and North Pacific [83], East China Sea off Southern Zhejiang Province [84], and Alaskan Arctic [11].

Oxygenic photosynthesis is one of the most important biological processes on Earth as it provides the basis for O₂-dependent respiration [35, 85]. It evolved ∼2.4 billion years ago and led to a first large increase of atmospheric O₂. This increase is referred to as the Great Oxidation Event (GOE) that fundamentally changed the Earth’s atmosphere from an essentially anoxic to a moderately toxic state, which formed the fundamental prerequisite for O₂-dependent metabolism [86]. It is hypothesized that oxygenic photosynthesis was initially performed by ancestral cyanobacteria [87] that provided for the increase of O₂ in atmosphere and oceans thereby paving the way for the development of complex life. Around 800‒600 million years ago, a second oxygenation event significantly increased the atmospheric O₂ concentration on Earth to similar levels prevailing today [88]. This event was termed Neoproterozoic Oxygen Event (NOE) and coincided with the origin of marine microalgae [89].

Microalgae have a complex evolutionary history as they have integrated genes from photosynthetic organisms, heterotrophic eukaroytes, and bacteria [90]. Beside general evolutionary principles, e.g., mutation, selection or gene drift, microalgal evolution was also driven by endosymbiotic events or vertical and horizontal gene transfer [90]. For example, green and red algae were integrated as chloroplasts via secondary endosymbiosis [91‒93]. Subsequent to this integration, microalgae separated into groups such as euglenoids and dinoflagellates, which distinguish themselves by green and red chloroplasts, respectively [90]. Interestingly, some dinoflagellates further engulfed chlorophyta in serial secondary endosymbiosis as well as haptophytes, cryptophytes, or diatoms in tertiary endosymbiotic events [94‒96]. These multiple evolutionary processes led to the tremendous biodiversity of current dinoflagellates, with significant increases in abundance and diversity observed during the consecutive geologic periods of Jurassic (208–144 million years ago) and Cretaceous (144–66 million years ago) [97‒102].

The current phylogenetic tree of eukaryotes classifies dinoflagellates into the so-called TSAR subgroup, which consists of the Telonemia, Stramenopiles, Alveolata, and Rhizaria (Fig. 3a) [103]. Within the Alveolata, dinoflagellates form a monophyletic group, alongside the ciliates, apicomplexans, and perkinsozoa (Fig. 3b). Consistent validation of their monophyletic clustering is based on various phylogenetic studies [101, 104]. Dinoflagellates display various lifestyles, ranging from free-living via endosymbiotic to parasitic (Fig. 3b; online suppl. Table S2). The order Prorocentrales, which harbors P. cordatum, belongs to the thecate (plate-bearing) core-/dinoflagellates and are morphologically diverse (including their plate patterns) (Fig. 3b) [101, 104, 105]. The theca of dinoflagellates is suggested to have evolved from a single origin, which is marked by a filled circle in the phylogenetic tree displayed in Figure 3b [101]. Members of the genus Noctiluca represent the earliest branching core-/dinoflagellates [101, 106]. Analysis of single-cell transcriptomes of diverse members of the order Noctilucales suggested that the atypical biology has emerged more recently within the group [107]. By contrast, Oxyrrhis marina represents an early branch of the dinoflagellate lineage and is regarded as a so-called pre-dinoflagellate [108]. Likewise, outside of the core-/dinoflagellates, parasitic marine Alveolates (MALVs) have independently evolved from free-living ancestral predators, i.e., the eleftherios and the genus Oxyrrhis [109]. Single-cell transcriptomics of under-sampled heterotrophic taxa revealed repeated gain and loss of various functionalities, including histone-like proteins (HLPs) and rhodopsins, which add to the complexity of dinoflagellate evolution [110].

To date, dinoflagellates comprise ∼2,500 living species, which are classified into ∼300 genera [112]. Based on contemporary knowledge, the majority of dinoflagellates occur as free-living organisms and around half of the inventoried species perform photosynthesis. Taxonomic classification of dinoflagellates is an ongoing scientific debate since many species are still reclassified or even newly described. Taxonomy of dinoflagellates remains highly challenging because it predominantly relied on microscopic observations; biological and phylogenetic information remained unconsidered. To date, phylotranscriptomic and genomic approaches as well as advanced microscopic techniques support previous morphology-based classification and facilitate further discovery of new dinoflagellate species. The taxonomic lineage of P. cordatum (formerly Prorocentrum minimum) is outlined in Figure 3c.

Dinoflagellates catch the eye by their morphological variety and cell biological peculiarities, such as a striking cell covering (some with theca) and a highly atypical nucleus (dinokaryon) [104, 112‒115]. The most distinct morphological features of dinoflagellates are summarized in the following with a focus on the order Prorocentrales and the species P. cordatum. A schematic drawing, which illustrates the morphology and subcellular architecture of P. cordatum based on TEM images of longitudinal sections, is provided in Figure 4.

The size of most described dinoflagellate species ranges between 10 μm and 100 μm [112]. Examples of very small species are Peridinium inconspicuum, Prorocentrum spinulentum, and P. cordatum, with lengths of ∼10–13 µm, ∼9–13 µm, and ∼10–24 µm, respectively [116, 122, 123]. By contrast, the dinoflagellate Noctiluca scintillans is a very large representative with a cell length of ∼100–800 µm [124, 125]. The cell shape of dinoflagellates varies greatly, including unique features such as expanded U-shaped horns of the genus Tripos (formerly part of the genus Ceratium) [126‒130] or wing-like extensions of the cell wall of the genus Ornithocerus [131]. Beyond that, cell shapes can be variable within a single species as demonstrated by some members of the genus Prorocentrum [117]. Across species of the genus Prorocentrum morphological features beyond shape and size, such as presence/size of spines and pores, can vary profoundly [46].

The cell covering of dinoflagellates is termed amphiesma and can take very primitive (e.g., genus Amphidinium [132]) or highly complex shapes (e.g., genus Histoneis [133]). Central to the theca design is the layer of flattened amphiesmal vesicles (AVs), which are confined by an inner membrane (cytoplasmic membrane [CM] = inner amphiesmal vesical membrane) and an outer plate membrane (outer amphiesmal vesical membrane [OPM]) [105, 121, 134, 135]. Within the AVs, individual cellulosic thecal plates (CTPs) are synthesized in the case of armored (thecate-bearing) dinoflagellates [104, 121]. The plasma membrane (PM) represents the outermost boundary of the cell also enclosing the flagellum [121]. The CTPs in turn can be highly diverse with respect to number, size, thickness, structure, and orientation around the cell [4, 112, 113]. The primitive cell covering (naked dinoflagellates; e.g., genus Oxyrrhis) is characterized by irregularly arranged AVs without CTPs [113]. By contrast, members of the genus Heterocapsa possess vesicles with numerous plates, together forming a sturdy theca [113, 136]. The order Prorocentrales harbors a simple type of cell covering including the following major features of the theca: (i) reduced to two large, ridge-bearing plates, which are complemented by ∼8–10 small platelets in the periflagellar area (Fig. 4a) [113, 117, 118], (ii) covered with thecal-enclosed spines and randomly distributed trichocyst/mucocyst pores (Fig. 4a) [117], and (iii) in some cases, massive cellulosic reinforcement, which provides the cells with a robust shelter against surrounding environmental forces and possibly ensures their survival under harsh conditions. The thecal pores of P. cordatum are erratically distributed without the formation of a specific pattern [116]. Structures of spines and pores were resolved in detail by atomic-force microscopy for P. donhangiense [137]. During growth/cell cycle and in response to various external stressors, the cell wall of thecate-bearing dinoflagellates (including P. cordatum) is massively rearranged in a process termed ecdysis; this involves shedding of the theca and flagellum yielding the formation of immotile ecdysal cysts [121, 138, 139].

Furthermore, some Prorocentrum species are highly variable in cell shape as they can be circular, oval, cordiform, or triangular and each species is characterized by an individual thecal plate surface (Fig. 4b) [117]. These characteristics include in particular the thecal ornamentation and pore pattern as well as the periflagellar pore area (Fig. 4b) [116‒118]. One of the most extraordinary and diverse cell shapes among dinoflagellates is known from members of the genus Tripos as their theca differ in direction and length of expanded horns and ornamentation [6].

Dinoflagellates typically possess a longitudinal and a transverse flagellum, which differ in their external morphology and extra-axonemal structure. The inner structure of the two flagella shows the typical “9 + 2”–arrangement of microtubuli. The longitudinal flagellum is characterized by short hairs and a large volume of packing material, whereas the transverse flagellum has unilateral, long, fine hairs, and a wavy ribbon-like structure [113, 140].

Moreover, dinokont, desmokont, and opisthokont flagellation can be distinguished, depending on the flagellar insertion site on the cell surface and in relation to the transverse groove [119]. Most described dinoflagellate species have a dinokont flagellation (e.g., order Peridiniales), where the longitudinal flagellum inserts in the posterior plane and the transverse flagellum beats in the girdle around the cell [113, 119, 140]. The desmokont flagellation could be observed for the order Prorocentrales, where both flagella are not associated with furrows and arise in anterior position [140]. The opisthokont flagellation is typical for the genus Oxyrrhis and is characterized by two posterior flagella distinct from those of core-/dinoflagellates [113].

In P. cordatum, the longitudinal and transverse flagella penetrate the periflagellar area via the flagellar pore (fp), which is surrounded by multiple small, differently sized platelets (Fig. 4a) [118]. A general concept on the structure of the flagellar apparatus of dinoflagellates has been depicted by Okamoto and Keeling for predatory Psammosa pacifica [119, 141] (Fig. 4c, left panel), which represents an early branch of the dinoflagellate lineage. Furthermore, EM studies revealed the spatial interaction of the flagellar apparatus with the pore canal in the case of P. cordatum and Prorocentrum micans (Fig. 4c, right panel) [120, 142]. Essentially, the scheme for P. pacifica (Fig. 4c, left panel) shows that the two flagella are connected to the longitudinal (bb1) and transverse (bb2) bodies, respectively. These two basal bodies are arranged at probably species-specific angles to one another (mostly 90°). Bb1 and bb2 are surrounded by a longitudinal and a transverse collar (LSC and TSC), which in turn are directly connected via a striated collar connective (SCc). Furthermore, the basal bodies are tightly associated with various microtubular roots and strands, giving rise to an overall complex structure [119]. Despite these structural insights into the flagellar apparatus, a mechanistic understanding of the flagellar movement is yet to be achieved.

Around half of the described dinoflagellate species are phototrophs harboring chloroplasts for performing oxygenic photosynthesis. Dinoflagellates have acquired chloroplasts via secondary or tertiary endosymbiosis by engulfing red- and other microalgae [91, 94]. The ancestral algal endosymbiosis was accompanied by the integration of the pigment peridinin leading to the so-called peridinin-dinoflagellate plastid, originating from red algae [143]. These special plastids also stand out for their unusual plastid genome with respect to loss of ancestral genes and unusual organization [98, 144]. Early electron microscopic studies indicated that the chloroplasts of dinoflagellates are structurally similar to those of other microalgae except for a specific envelope [113]. The latter is mostly made of three membranes, which can be wrinkled to some extent and have a combined mean thickness of ∼230 Å [143, 145]. A chloroplast girdle lamella as known from diatoms is lacking in dinoflagellates [146].

The architecture and orientation of chloroplasts in dinoflagellates are highly variable. For example, they are saucer-shaped, located at the cell’s periphery, and occupy ∼30% of the cell volume in small, marine Acrodontium pigmentosum (formerly Aureodinium pigmentosum) [147], or are more lens-shaped and reticular in bloom-forming Alexandrium tamarense (formerly Gonyaulax tamarensis) [145, 148].

The chloroplast contains numerous lamellae running in parallel, which are mostly composed of three apposed thylakoid membranes but in some cases also of up to 30 in a single stack [145]. A grana-like arrangement of thylakoid membranes has not been reported for dinoflagellates. The stroma of dinoflagellate chloroplasts contains granular materials and lacks intersecting lamellae in some regions, which are suggested to expand into pyrenoids. Under high light conditions, the lamellae are further apart from each other, forming a larger stromal area [145]. The chloroplast stroma contains evenly dispersed ribosomes and DNA in the form of irregularly arranged fibrils [145].

Previous TEM-based images of P. cordatum cells suggested similar chloroplast structures and arrangement of the thylakoid membranes; however, the number of chloroplasts per cell remained unclear (Fig. 4d, 5a‒d) [113, 145]. A recent focused ion beam/scanning electron microscopy (FIB/SEM)-based 3D reconstruction of the chloroplast of P. cordatum revealed a single, reticulate, barrel-like chloroplast, which completely lines the inner cell envelope and occupies ∼40% of the cell volume (Fig. 5e, f) [149]. Reticulate chloroplasts, occupying similar shares of the cell volume, were also reported for other dinoflagellates and microalgae and suggested to be formed under certain environmental conditions (e.g., lifestyle or light regime) [150‒153].

Pyrenoids are special regions within the chloroplast for enhanced carbon-concentrating and CO₂-fixation [155]. They appear as dense structures in EM images, lack membranes and usually localize inside the chloroplast or tightly associate with it [144, 156]. Dodge and Crawford [144] classified pyrenoids into five categories: (i) simple interlamellar, (ii) compound interlamellar, (iii) single-stalked, (iv) multi-stalked, and (v) stalked with invaginations.

They contain high amounts (up to 90%) of ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) and associate with starch reflecting their central role in photosynthetic CO₂-concentration and CO₂-fixation [156‒159]. Carbon-concentrating mechanisms (CCM) in conjunction with high RuBisCO amounts have evolved to compensate for the low catalytic rate of this enzyme [155, 160, 161]. Concentrating of RuBisCO in the pyrenoid without involvement of membranes is achieved by liquid-liquid phase separation (LLPS) yielding droplets with high protein concentration surrounded by a dilute phase. This arrangement facilitates higher diffusion rates between the phases than achievable via membrane-enclosed compartmentalization. Multiple hexadecameric assemblies of RuBisCO are tethered via the filamentous EPYC1 protein, yielding the high concentration of the enzyme in the pyrenoid [154, 162, 163]. The ultrastructure and the energetic mechanism of pyrenoids are to date best understood for the unicellular, freshwater-inhabiting green alga Chlamydomonas reinhardtii (Fig. 5g) [154, 164]. Metabolic exchange between the pyrenoid matrix and the chloroplast stroma is mediated by pyrenoid tubules, which are formed by penetrating thylakoids. Thereby, delivery of CO₂ to RuBisCO and in the reverse direction of generated 3-phosphoglycerate to the Calvin Cycle is rendered possible. Uptake of external HCO₃^‒ across the various membranes of the algal cell and its ultimate conversion to CO₂ are proposed to be energized jointly by mitochondria and chloroplast [164].

Pyrenoids are found in chloroplasts of almost all microalgae [165‒167], as well as in a single genus of vascular embryophytes (Anthoceros) and hornworts (sister group of vascular plants) [159, 168], while they are absent in the remaining terrestrial plants. Notably, hornworts have repeatedly gained and lost pyrenoids during the last hundred million years [169]. This argues against the view that the origin of pyrenoids was linked to low atmospheric CO₂ levels [170].

The order Prorocentrales has long been known to possess large, internal chloroplasts. These contain pyrenoids, which can be differently positioned within the cell and were previously reported to have the following major features [144, 171]: (i) two large pyrenoids located midway between apex (anterior) and base (posterior) on both sides of the sagittal axis, (ii) swollen chloroplast regions near the pyrenoids often projecting towards the cell’s interior, and (iii) high amounts of aggregated protein, contrasted by the absence of starch [144]. These early observations were corroborated at high resolution in the recent 3D reconstruction of the P. cordatum chloroplast (Fig. 4d, 5e, f) [149].

Mitochondria of dinoflagellates are structurally similar to those of land plants or higher eukaryotes and possess envelopes, which are composed of two membranes [172]. Mitochondria of dinoflagellates are often elongated, oval-shaped, or form irregular structures [113]. Their lumen is tightly packed with the inner membrane, harboring the typical, slightly constricted, tubular cristae [172]. TEM-based images of P. cordatum revealed the presence of numerous mitochondria dispersed in the cytoplasm (Fig. 4d, 6a, b). However, most recent 3D reconstructions of mitochondria from diverse microalgae showed that they form large, reticular networks, occupying ∼2.5–5% of the cell volume as exemplified for P. cordatum in Figure 6c, d [149, 150, 152, 153]. It was also demonstrated that mitochondria are ∼two-fold more voluminous under conditions of increasing light [151]. However, structural and molecular biological studies on the mitochondria in dinoflagellates are scarce, with the process of oxidative phosphorylation (OXPHOS) localized there also under-investigated.

At present, it is unclear whether the observed reticular structure of a single, large mitochondrion represents a permanent cellular feature of P. cordatum. In this context, it is noteworthy that mitochondria of human cells are subject to structural dynamics, e.g., during cell autophagy and apoptosis [172‒175]. These cellular responses involve a plethora of proteins and induce the formation of large, reticular networks by mitochondrial fusion and fission events. In parts, these reticular mitochondria form interconnected networks, e.g., with the nucleus to facilitate the transport of ATP and metabolites.

The nucleus of dinoflagellates, also termed dinokaryon, stands out by its multiple exceptional features profoundly differing from the nuclei of other eukaryotes. These features include a very high content of nuclear DNA, which occurs in a crystalline state and is divided up into many condensed chromosomes, as well as an unusual mitosis and diverse extraordinary proteins. In the following, current knowledge about the nuclear envelope (NE) and its pores, the properties of chromosomes, and the mechanism of nuclear division are summarized.

The NE is made up of two membranes, which have a combined thickness of 7–8 nm, is partly interconnected with the endoplasmic reticulum, is connected via a lamina to the cytoskeleton, and is mostly perforated by numerous nuclear pores [176, 177]. These nuclear pore complexes (NPCs) are macromolecular assemblies that control the exchange of (macro)molecules between nucleus and cytoplasm [178]. Knowledge about the number per nucleus, structure, and composition of NPCs in dinoflagellates is very limited. The arrangement of nuclear pores varies from randomly distributed in Kryptoperidinium (formerly Glenodinium) to hexagonally packed in Prorocentrum [96, 176, 179]. However, in some dinoflagellates NPCs seem to be absent and are probably substituted by / integrated in vesicle-like structures (nuclear chambers) attached to the NE, as inferred from EM studies with Noctiluca and Gymnodinium [113, 180‒182]. A recent FIB/SEM-based 3D reconstruction of the NE of P. cordatum revealed 475 NPCs, which are to the largest part arranged in a confined area close to the nucleolus [115]. This patch-like arrangement presumably facilitates interactions and transport processes between the nucleolus, as active hotspot of the nucleus, and other organelles [115, 183]. Notably, proteomic analysis of the nuclei-enriched fraction of P. cordatum revealed only few homologs to known NPC protein components of well-studied yeast Saccharomyces cerevisiae [184] but rather a multitude of proteins of unknown function, implicating a fundamentally different composition and possibly architecture [115].

Dinoflagellates often organize their considerable DNA content in a large number of chromosomes, which are in a liquid-crystalline state, permanently condensed throughout the cell cycle, and lacking typical nucleosomes [185, 186]. For half a century, numerous studies addressed the architecture of these conspicuous chromosomes yielding various models [187‒198], which can be summarized as follows.

The packed DNA exists in a cholesteric-like state, coining the term liquid-crystalline chromosomes. The term cholesteric describes a specific type of liquid-crystalline state, which is characterized by a helical molecular arrangement. This represents the highest degree of DNA condensation and has so far been detected exclusively in dinoflagellates and animal sperm cells [199, 200]. Early EM images revealed that the chromosomes are covered with a series of transverse bands (fibril-like structures) and are apparently composed of multiple stacks of DNA discs twisted against each other [113, 191], which likely reflects the helicoidal arrangement of cholesteric-like DNA. The structural organization, condensation, and stabilization of the chromosomes are apparently mediated by small proteins and high amounts of bivalent cations, which neutralize the negative charge of the DNA backbone [197]. These chromosomal proteins include, e.g., HLPs or dinoflagellate viral nuclear proteins [201‒203]. Phylogenetic analysis suggests a paraphyletic origin of DNA-binding proteins in dinoflagellates’ nuclei comprising three evolutionary roots [101].

A certain structural heterogeneity of chromosomes derives from a highly condensed core contrasted by peripheral loop regions, which are regarded as the sites of transcriptional activity and of a given stage of replication [204, 205]. At present, it is unclear how chromosomal architecture featuring a central protein-rich/DNA-poor core fiber, wrapped by a DNA/cation-rich chromonema (Fig. 7a) [197] relates to the previous concept of cholesteric condensed DNA ordered in discs twisted around each other (Fig. 7b, c) [206]. The number of chromosomes varies considerably among different dinoflagellate species and range from a few to several hundred in number [207].

In the case of P. cordatum, recent TEM images underpin the fibril-like structuring of the chromosomes as well as the large spatial extent of the nucleus (Fig. 7d, e). A subsequent, detailed FIB/SEM analysis revealed a total number of 62 tightly packed chromosomes (occupying 80% of the nuclear volume) featuring markedly different sizes and volumes (Fig. 7f, g) [115, 149]. Further observations from the 3D reconstruction concerning the chromosomes were: (i) v-shape in some cases as an indicator of ongoing replication, (ii) attachment to the NE presumably in the context of dinomitosis, and (iii) interaction with the nucleolus suggesting ongoing transcription. Furthermore, proteomic analysis of enriched nuclei from P. cordatum identified major basic nuclear proteins (MBNP; synonym: HLPs) as the most abundant proteins interacting with the chromosomes. These very small (∼10 kDa) MBNPs are known to have a weak DNA-binding ability [208], predominantly localize at the chromosomal periphery, and are suggested to be involved in transcription [205].

Dinoflagellates orchestrate a type of nuclear division in a unique way, termed dinomitosis. Its characteristic feature is the closed mitosis, where the NE stays intact and the chromosomes remain permanently condensed [194, 209, 210]. Dinomitosis was thus far investigated with several dinoflagellates, such as Crypthecodinium cohnii [211, 212], Oodinium fritillariae [213], Amphidinium carterae [190, 214], P. cordatum [215], Polykrikos kofoidii [209], and Ostreospis cf. ovata [210], providing first insights into this fascinating cell biological process, with a molecular-mechanistic understanding still remaining rudimentary.

Barlow and Triemer [214] proposed a model of the mitotic spindle apparatus of A. carterae and suggested how nuclear division may be performed in dinoflagellates (Fig. 8a). The process starts with the invagination of the NE from both nuclear poles forming one or more trans-nuclear tunnels (spindle channels), the interior of which is continuous with the cytoplasm and spatially separated from the nucleoplasm by the NE. Each tunnel accommodates a spindle consisting of 2–4 bundles of nonchromosomal microtubules, which are connected to each spindle pole forming two interdigitated half-spindles. The spindle apparatus is completed by NE-embedded kinetochore microtubules. These bind to the nucleoplasm-localized chromosomes as well as the tunnel-residing spindle, thereby establishing an NE-bridging link between both of them. Then, elongation of the spindle leads to an expansion of the nucleus, while maintaining a constant position of the separating chromosomes relative to the nuclear poles. When the nucleus finally divides, the spindle disassembles and the tunnels collapse, yielding two spherical daughter nuclei. At present, it is unclear how DNA replication and nuclear fission are coordinated and whether the permanent linkage of the spindle apparatus with the endoplasmic reticulum controls nuclear division.

A recent microscopic study of dinomitosis in the predatory dinoflagellate P. kofoidii identified six trans-nuclear tunnels interconnected with further membranous extensions forming the so-called nuclear net [209]. This endomembrane network consists of proteinaceous structures and extends (branches out) across the entire nucleus. While the function of this nuclear net remains elusive at present, it was also observed in dividing nuclei of P. cordatum (Fig. 8b) [115, 215], C. cohnii [211, 212], and Kryptoperidinium triquetrum (basionym: K. foliaceum) [216, 217]. Most recently, intricate microtubule structures maintained throughout dinomitosis were described for free-living, benthic Ostreospis cf. ovata [210], further demonstrating the unique cellular features of nuclear division in dinoflagellates.

Golgi bodies or dictyosomes are present in almost all microalgae and are generally involved in protein modification, processing, and packing. While the Golgi apparatus of dinoflagellates is rarely studied, EM studies, nevertheless, showed that it consists of multiple flattened, membrane-bound cisternae, and attached vesicles [113]. Thus, it is structurally similar to the Golgi apparatus of other eukaryotes. The filigree membrane structure of the Golgi apparatus of P. cordatum, as resolved by TEM (Fig. 9a) and FIB/SEM (Fig. 9b) analyses [149], agrees with previous findings [218, 219]. Furthermore, the 3D reconstruction of an entire cell of P. cordatum revealed that the Golgi apparatus occupies 0.3% of the total cell volume and tightly intercalates with the single, reticular mitochondrion [149].

Dinoflagellates store large amounts of biomacromolecules in vesicle-like structures or granules, which play significant roles in the maintenance of the cell’s metabolic homeostasis as well as the dynamics of nutrient availability in the habitat [113, 158, 220, 221]. Moreover, these storage inclusions also bear broad biotechnological potential. The stored biomacromolecules include polysaccharides, lipids, and polyphosphates, which can be distinguished in EM images on the basis of their distinct electron densities according to Dodge [113] (Fig. 9c). It can be assumed that storage inclusions are formed under conditions of nutrient surplus and re-metabolized upon nutrient limitation viz. starvation. Notably, dinoflagellates also store large amounts of protein (RuBisCO) in the chloroplast-associated pyrenoids (see above).

Microalgae store polysaccharides (mostly starch) in variably sized, grain-like structures (starch grains), which can take a flattened, oval, spherical, or dome-shaped form [113]. This early study by Dodge [113] revealed organism-specific locations of these starch grains in the cell: (i) within the chloroplast as in Chlorophyceae or Prasinophyceae, (ii) randomly distributed in the cytoplasm as in Dinophyceae and Rhodophyceae, or (iii) as large droplets at the posterior end of the cell as in Chrysophyceae or Haptophyceae. A previous study [167] with marine and freshwater dinoflagellates reported the starch grains to be more abundant during daytime and located in proximity to the chloroplast. Notably, in P. cordatum, multiple starch grains, accounting for ∼1.8% of the cell volume, are distributed within the single, barrel-shaped chloroplast (Fig. 9c, d) [149].

Microalgae generally store lipids as membrane-free droplets of variable form and size in the cytoplasm, which are readily recognized in EM images [222, 223]. In the dinoflagellates A. carterae, Cystodinium sp., and Apocalathium aciculiferum (formerly Peridinium aciculiferum) long polyunsaturated fatty acids were identified by high resolution mass spectrometry [224]. Dinoflagellates from the North Pacific were shown to intermediately synthesize energy-rich lipids during the day to subsequently consume them during night [225]. P. cordatum grown under the applied conditions produced multiple lipid droplets, accounting for ∼3.5% of the cell volume, as determined by FIB/SEM-based 3D reconstruction of the cell (Fig. 9c, e) [149].

Microalgae as well as cyanobacteria form and use inorganic polyphosphates for a broad range of purposes including provision of phosphates for anabolism (in particular (poly)nucleotides and phospholipids), cation homeostasis, and diverse stress responses [220, 226, 227]. Formation of polyphosphates is usually promoted by environmental stress including nutrient limitation or imbalance [228]. In the green alga Trebouxia sp., the number of polyphosphate inclusions decreased under phosphate limitation [229]. Polyphosphates of dinoflagellates appear as electron-dense structures in electron microscopy, representing spherical granules [113]. In a 3D reconstructed cell of P. cordatum, numerous polyphosphate granules, accounting for ∼1.5% of the cell volume, are distributed randomly in the cytoplasm (Fig. 9c, f) [149].

Dinoflagellates possess various types of extrusomes such as mucocysts, nematocysts, or trichocysts [113, 230]. In general, extrusomes are cellular, membrane-enclosed, fibrous structures observed in protists, which can be ejected by the cell in response to environmental triggers [231‒233]. Mostly, cells discharge these structures to escape predators or to defend themselves against possible predators [232, 234]. The most prominent type of extrusomes in dinoflagellates is represented by the trichocysts, which strongly resemble the spindle trichocysts of ciliates [235]. After their synthesis in the Golgi apparatus, trichocysts reside on the cytoplasmic side of the cell membrane, perpendicularly positioned to the cell surface [235]. These non-elongated (or resting) trichocysts are composed of proteinaceous rods, tubular elements, and tubular fibers [235]. Their core is a dense, crystalline, rhombic or square-like structure, readily detectable by electron microscopy [235]. Upon being triggered, trichocysts are elongated up to 200 μm and discharged through the pores of the thecal plate [235]. The fine structure of elongated trichocysts displays a regular, periodic striation, which is different among dinoflagellates [113, 235‒237]. In the case of P. cordatum, ejected long trichocysts as well as their exit pores in the theca could be detected by scanning electron microscopy (Fig. 10a, b). TEM images provide various perspectives of membrane-enclosure as well as rhombic structure of trichocysts resting within the cytoplasm of P. cordatum (Fig. 10c–f).

Several operating modi are considered for extrusomes to discharge their specific contents. (i) In the case of two dinoflagellates, marine Nematodinium sp. and predatory P. kofoidii, FIB-SEM analysis of their nematocysts indicated two disparate modes of “ballistic” discharge, i.e., simultaneous expulsion of about a dozen projectiles versus a single projectile released from a pressurized capsule (Fig. 10g, left and middle panel) [238]. (ii) Upon protein-mediated specific cell-cell interactions, the obligate intracellular malaria parasite Plasmodium falciparum, belonging to the Apicomplexa (sister group of dinoflagellates; Fig. 3b), injects a protein cocktail through apical rings of the rhoptry (secretory organelle) into the host cytoplasm to facilitate invasion [240‒242]. (iii) The freshwater ciliate Paramecium ejects trichocysts apparently through a similar ring-enclosed rosette [240, 243]. Most recent cryo-electron microscopy-based studies with “ballistic” discharging Nematodinium sp. and P. kofoidii [238] as well as the rhoptry discharging Toxoplasmo gondii, a parasitic member of the Apicomplexa [239], provided unprecedented insights into the molecular architecture and expulsion mechanism of extrusomes as depicted in Figure 10g (right panel).

The hallmarks of dinoflagellate genomes are their extremely large size and various atypical features such as high numbers of dispersed duplicates, multi-codon genes, or a polycistronic transcript structure [30, 244‒246]. The most distinct genomic features of dinoflagellates are summarized in the following with focus on the species P. cordatum.

Haploid genomes of dinoflagellates vary greatly among the species and are much larger compared to those of other microalgae [247]. Dinoflagellate genomes are estimated to range from ∼1 to 250 Gbp, while genomes of studied diatoms, haptophytes, and prasinophytes are a ∼1,000-fold smaller with sizes of ∼12–165 Mbp [207, 247, 248]. In addition, compared to the human genome, dinoflagellates possess up to 70-fold larger genomes, contradicting a direct correlation between genome size and complexity of the organism. However, the lifestyle of dinoflagellates apparently governs the genome size as parasites and endosymbionts harbor markedly smaller ones than free-living species (Table 1). This genomic downsizing is attributed to the process of reductive evolution, which streamlines the genomes of endosymbionts and parasites in response to the stable and nutrient-rich host environment [249‒252]. A recent study of the free-living dinoflagellate P. cordatum revealed a large, haploid genome of 4.15 Gbp [30]. Compared to other dinoflagellates, it represents a rather large genome, which is ∼4 times larger than those of endosymbiotic dinoflagellates (mean ∼1 Gbp) (Table 1) [24, 30, 253‒261].

The extremely large size of dinoflagellate genomes is hypothesized to relate to their complex gene structure. The genomes harbor a high number of protein-coding genes (mean of ∼50,000; Table 1), with the majority of them existing as duplicates [262]. They are probably integrated into the genome by very numerous duplication and recombination events yielding multiple gene copies sequentially arranged as so-called tandem repeats in the DNA [263]. The key enzyme RuBisCO, e.g., is encoded in triple tandem (in Symbiodiniaceae) or quadrupole tandem (in Prorocentrum) repeats [264, 265]. Another peculiarity is that some dinoflagellate gene models are composed of multi-codon units (CUs), which share high sequence identities and similarities with each other and with CUs of paralogous gene models (Fig. 11) [30, 149, 246]; for definition of the term gene model refer to Gerstein et al. [266] and Schnable [267]. The genome of P. cordatum possesses gene models for pigment-binding proteins (PBPs), which contain up to 12 chlorophyll-binding domains (i.e., CUs), sharing 70–100% sequence similarity among each other (Fig. 11a). Such an arrangement is further exemplified for major basic nuclear proteins (Fig. 11b) and RuBisCO (Fig. 11c) [149]. Furthermore, dinoflagellates express so-called polycistronic transcripts, which are defined as single mRNAs encoding two or more peptides or proteins [245]. These transcripts are suggested to be converted into monocistronic sequences by trans-splicing of a dinoflagellate-specific spliced leader sequence (dinoSL) [268]. Finally, the DNA of dinoflagellates is highly methylated, and the degree of methylation is suggested to depend on the experienced light conditions [269]. Highly methylated DNA is suggested to be involved in modifying chromosome structure and regulating gene expression [270, 271].

P. cordatum possesses 85,849 protein-coding genes, of which 64% (∼54,700 genes) are duplicates. More than 70% of the protein-coding genes are of unknown function: ∼50% of them are unique and most of the remaining ∼22% are matching to proteins previously predicted in other dinoflagellate genomes [30]. Additionally, P. cordatum has a higher GC-content (∼60%) [30] compared to other dinoflagellates and eukaryotes (∼40–55%) [253, 263, 272], which probably allows P. cordatum to survive under increasing temperatures [273]. Furthermore, the genome of P. cordatum revealed unusually large intron versus exon areas and numerous introner elements, suggesting a highly active transposition that establishes abilities to adapt to changing environmental conditions [30, 274].

Chloroplast genomes of dinoflagellates are highly reduced, particularly in species with peridinin-containing chloroplasts, and all of them are characterized by genes that are spread over so-called minicircles featuring species-specific sizes [275‒277]. In plants or other eukaryotic algae, chloroplast genomes harbor ∼100–250 genes, whereas chloroplast genomes of dinoflagellates are composed of only 16 genes [276]. The majority of these encode proteins involved in photosynthesis, such as constituents of the photosystem, the chloroplast ATP synthase, or the cytochrome b₆f complex [275]. These genes are located in sets of 0–4 on minicircles [278]. Their expression is suggested to be performed by a rolling-circle mechanism yielding long transcripts [276]. Posttranscriptional modifications of chloroplast transcripts are generally rare; however, 3′-polyuridinylation occurs in some species [276, 279].

The mitochondrial genome of dinoflagellates seems to be even more reduced, compared to that of chloroplasts, and codes for only cytochrome b, two cytochrome oxidases (Cox1, Cox3), and two fragmented rRNAs [280, 281]. Similar to the genes of the nuclear genome, the mitochondrial genes can exist as multiple duplicates with up to 10 copies per gene [282]. Dinoflagellates, like other apicomplexans, import tRNA from the cytoplasm to execute translation [280]. Interestingly, a recent study on the mitochondrial genome of the parasitic dinoflagellate Amoebophyra ceratii (a relative of the core-/dinoflagellates) revealed an active mitochondrion that apparently lacks any mitochondrial genes [253]. The chloroplast and mitochondrial genomes of P. cordatum have not been investigated so far; however, similar genetic structures as described above can be assumed.

DNA replication in dinoflagellates is a particularly demanding process due to the highly condensed cholesteric, liquid-crystalline chromosomes, where semi-ordered phases of DNA are assumed to facilitate separation of parental and daughter DNAs [206, 283]. On the molecular level, DNA replication is suggested to resemble that in other eukaryotes. This process involves several proteins, which could be identified by transcriptome and proteome data or at least predicted from the genomes of the dinoflagellates P. cordatum [115] and Lingulodinium polyedra [284]. These proteins include subunits of the DNA polymerase, DNA helicase, DNA ligases, primase, and the topoisomerase, as well as single-strand binding proteins, cell division control proteins, proliferating cell nuclear antigen proteins, replication factors of the clamp-loader class, and constituents of the origin of recognition complex. Further details on the replication machinery of P. cordatum are depicted in Figure 6a in Kalvelage et al. [115].

The process of transcription in dinoflagellates can be expected to differ fundamentally from that of other eukaryotes due to the aforementioned highly condensed nature of the DNA as well as absence of most proteins known to be involved in this process. The current state of knowledge is summarized in the following.

Dinoflagellates possess three different types of DNA-dependent RNA polymerases similar to other eukaryotes, where most core elements could be confirmed by transcriptomics [285]. Most of the common transcriptional factors (TFIIH, TFIIA, TFIIB, TFIIE, and TFIIF), which activate and control transcription, are, however, absent in dinoflagellates or show low similarities with those of other eukaryotes [263]. Nevertheless, most of the studied dinoflagellates share the transcriptional factor TFIIH, which is well known to be involved in regulating transcription and in nucleotide excision repair (NER) of damaged DNA [263, 286]. Interestingly, dinoflagellates possess large numbers of cold-shock proteins (CSPs) of unknown function, which were annotated as transcriptional factors and contain RNA-binding motifs [263, 287]. Furthermore, detailed investigations of CSPs in dinoflagellates of the genera Lingulodinium and Symbiodinium implicate their involvement in transcription [288, 289]. The TATA-binding protein (TBP), which binds to the promotor region, is generally absent in dinoflagellates [115, 290]. However, in C. cohnii a TBP-like protein (TFL), lacking the typical binding residues to attach to the TATA-box, was purified and demonstrated to bind a TTTT-box rather than the canonical TATA-box [290]. Subsequent studies corroborated the absence of the TATA-box from dinoflagellate genomes, which is substituted by the TTTT-box as transcriptional start site [263, 291].

P. cordatum lacks most proteins involved in transcription; however, around half of the components of the RNA polymerase, two transcriptional factors (TFIIB and TFIIH) and numerous CSPs could be identified by proteogenomics [115]. These findings agree with previous studies on dinoflagellate transcription and suggest that dinoflagellates probably have specific transcriptional features such as a distinct promoter region and recognition mechanism. Further details on transcription in P. cordatum are depicted in Figure 6d in Kalvelage et al. [115].

The process of translation is suggested to be highly conserved among eukaryotes, including dinoflagellates, occurs within the ribosomes and involves several initiation, elongation and other factors [292]. However, in the case of dinoflagellates several distinct features of translation have been reported thus far [288, 291]. The large ribosomal subunit is apparently smaller than typical of eukaryotes and ribosomal RNAs can contain specific nucleotides [293, 294]. In the case of eukaryotic translational elongation/initiation factors, several dinoflagellate-specific features have been reported: (i) eIF4E revealed a different cap-structure, (ii) eIF4G is apparently shorter, while eIF2 is similarly sized compared to other eukaryotes [292]. The RNA-binding proteins (RBPs) in dinoflagellates are poorly understood and show low sequence similarities with other RBPs [292, 295].

For P. cordatum various ribosomal proteins, translation initiation factors, elongation factors, RNA-binding proteins as well as methionine-tRNA ligases could be predicted and identified by proteogenomics [115]. Further details on the initiation of translation of P. cordatum are depicted in online supplementary Figure 5iii–iv in Kalvelage et al. [115].

The complex gene structures of dinoflagellates described above strongly challenge the identification of proteins and thus the interpretation of proteogenomic data, as recently shown for P. cordatum [30, 115, 149]. The difficulties encountered mainly arise from the high sequence similarities between homologous proteins: (i) The determined peptides often do not specifically match to a single predicted protein but are rather indistinguishable to all related homologs. (ii) Mapping of a specific peptide species to one member of a group of homologs could entail an erroneous co-mapping of all other related peptide species to this particular protein rather than to the actual cognate members. Another challenge is caused by highly abundant proteins (e.g., PBPs), which yield small and well-ionizable peptides. They dominate the MS-detectable proteins and thereby hamper the detection of proteins, which are of low abundance and beyond that yield peptides escaping MS analysis.

Photosynthesis encompasses diverse photochemical and enzymatic reactions to generate NADPH and ATP for driving CO₂-fixation as the basis for biomolecule synthesis. As a by-product of oxygenic photosynthesis, molecular oxygen (O₂) is released into the atmosphere or is further used for respiratory energy conservation of the cell, viz. OXPHOS in the mitochondria. The following section summarizes the major steps in these processes for P. cordatum, as recently reconstructed on the basis of multi-omics data (Figs. 12‒15) [30].

The light reaction of photosynthesis takes place in the thylakoid membranes of the chloroplast, where large, membrane-embedded protein complexes perform the photochemical reactions combined with a linked electron transport chain (Fig. 12a, upper panel) [30, 296]. This photosynthetic machinery is generally composed of light-harvesting complexes (LHCs), photosystem I (PSI), photosystem II (PSII), the cytochrome b₆f complex (cytb₆f complex), and a chloroplast ATP synthase (cATP syn) [297]. The major steps involved are as follows [296]. The pigment-bearing LHCs collect light energy, which they channel to the reaction centers of PSI and PSII. Here, light energy transforms the special chlorophylls into excited states at very negative redox potentials. Oxidation of the excited chlorophylls then leads to charge separation, i.e., oxidized special chlorophylls and reduced PS-specific primary electron acceptors. In the case of PSII, the “electron gap” in its special chlorophyll (Chl₆₈₀) is filled by cleavage of H₂O, yielding O₂ as by-product, and the electron released from excited Chl₆₈₀^* is transferred via the plastoquinone pool to the cytb₆f complex. In the case of PSI,the “electron gap” in its special chlorophyll (Chl₇₀₀) is filled by plastocyanin (receiving electrons from the cytb₆f complex) and the electrons released from excited Chl₇₀₀^* are transferred to ferredoxin. The latter is then oxidized by ferredoxin:NADP⁺ oxidoreductase furnishing NADPH. Water cleavage at PSII and one-way flux of electrons through the cytb₆f complex coupled to proton pumping [298, 299] generate the proton motif force (PMF), which in turn drives ATP synthesis via cATP syn. The architecture of this canonical machinery could be reconstructed in large parts for P. cordatum by proteogenomics (Fig. 12b). Notably, the detected protein components displayed increased abundances in response to heat stress, presumably to compensate for heat-induced metabolic disbalances [30], and form a conspicuous megacomplex (Fig. 12a, lower panel) [149].

Light-sensitive pigments absorb light by virtue of their delocalized π-electron systems. Pigments are bound to the LHC-proteins via specific or unspecific binding sites, together forming pigment-protein networks for relaying absorbed light energy to the reaction centers of the PSs. In some organisms, these networks can extend to giant complexes, which contain dozens of PBPs and hundreds of pigment molecules. In general, these pigments are classified as chlorophylls, carotenoids, or phycobilins.

Chlorophylls represent the most abundant pigment class on earth, absorbing blue-green and red light [300]. The basic structure of a chlorophyll molecule consists of a light-absorbing planar tetrapyrrole ring system, containing a central magnesium atom, and an attached hydrophobic poly-isoprenoid side chain for anchoring (Fig. 13a) [300]. The different types of chlorophylls (chlorophyll a‒f) are mainly defined by the substituents attached at various positions of the tetrapyrrole ring system [300].

Carotenoids are structurally highly diverse, widely distributed pigments that absorb yellow, orange, red, and purple light [301]. They serve as photoprotective agents, antioxidants, accessory light-harvesting pigments, and as precursor for plant hormones [302‒304]. They have a tetraterpene basic structure, which is characterized by a polyene chain with nine conjugated double bonds and a terminal group at both ends. Carotenoids are classified as carotenes or xanthophylls. The ∼50 different types of carotenes include, e.g., α-, β- and γ-carotene, and lycopene. Xanthophylls are markedly more diverse (∼800 different types) and comprise, e.g., β-cryptoxanthin, lutein, zeaxanthin, astaxanthin, fucoxanthin, and peridinin.

Phycobilins are mainly found in cyanobacteria or red algae and capture blue, red, and green light [305]. They function in light harvesting, as antioxidant, or as antibacterial and antitumor agent [306]. They have an open-chain tetrapyrrole basic structure and are sub-categorized as phycocyanin, phycoerythrin, or allophycocyanin [305].

Dinoflagellates harbor a great diversity of photosynthetic pigments. Zapata et al. [31] identified 63 different pigments from 64 dinoflagellate species based on HPLC analyses. The number and composition of pigments in dinoflagellates is controlled by a broad range of factors, including the type of habitat [307], seasonal dynamics [308], nutrient stress [309], and temperature, salinity, or growth phase [310]. In the case of P. cordatum, six different pigments (chlorophylls and carotenoids) were detected under the applied, controlled laboratory conditions [149], including chlorophyll a and c₂, violaxanthin, diadinoxanthin, peridinin, and β-carotene (Fig. 13a). The absorbance spectra of pigments of P. cordatum cover a broad light spectrum (Fig. 13b), which should enable survival under constantly changing light regimes and in deeper layers of the euphotic zones as described also for other photosynthetic microalgae [311]. Lastly, the P. cordatum pigments are readily separable and detectable by HPLC-analysis (Fig. 13c).

The dark reaction of photosynthesis is conducted in the stroma of the chloroplast, using atmospheric CO₂ (via the HCO₃^‒-equilibrium) and ATP/NADPH (from the light reaction) to generate C₃-precursors for anabolism [312]. CO₂-concentration facilitates CO₂-fixation by RuBisCO and can proceed via two principal mechanisms (CCMs). A widely used mode of the CCM is the uptake of HCO₃^‒ via ATP- or PMF-energized transporters followed by its dehydration to CO₂, which is catalyzed by carbonic anhydrase (CA), together increasing the CO₂ levels at the RuBisCO sites [160]. An alternative, efficient strategy for CCM is that used by C₄-photosynthesis (Fig. 14a) [313]. Here, PEP carboxylase initially fixes HCO₃^‒ to PEP, furnishing the C₄-dicarboxylate oxaloacetate (OAA), which is converted to malate by malate dehydrogenase. Then, NADP⁺-dependent malic enzyme oxidatively decarboxylates malate, forming pyruvate and CO₂, the latter of which is provided to RuBisCO. Use of C₄-photosynthesis also by the ubiquitously occurring marine diatom Thalassiosira weissflogii was demonstrated based on isotope (¹⁴C) and enzymatic analysis [314, 315], explaining the until then non-understood capacity of diatoms to avoid CO₂-limitation. More recently, C₄-photosynthesis was also reported for the green tide-forming marine macroalga Ulva prolifera [316] and the red tide-forming marine dinoflagellate Prorocentrum donghaiense [317], underpinning its relevance for marine phototrophs. The CO₂-fixation product 3-phosphoglycerate is subsequently converted to glyceraldehyde-3-phosphate (GAP) in the Calvin cycle (Fig. 14a) and channeled to either produce starch as carbon storage or to supply the multiple anabolic routes via the lower branch of glycolysis and TCA cycle (Fig. 14b). Most of GAP is, however, used in the Calvin cycle to regenerate the primary CO₂ acceptor ribulose-1,5-bisphosphate (RuBP). The outlined pathways were essentially reconstructed from the proteomic data of P. cordatum (Fig. 14c) and shown to have reduced protein abundances in response to heat stress, reflecting disturbed cellular homeostasis [30]. Furthermore, ample formation of starch by actively growing P. cordatum could be visualized in the 3D reconstructed chloroplast [149].

OXPHOS occurs in the inner mitochondrial membrane and involves the respiratory chain (Fig. 15a). OXPHOS is based on the transfer of electrons from NADH and FADH₂, generated in the central metabolism, to the final electron acceptor O₂ coupled to the formation of PMF, which drives ATP synthesis [318].

In general (including land plants such as Arabidopsis thaliana), the respiratory chain is composed of several membrane-embedded protein complexes (NADH-dehydrogenase complex, complex I; succinate dehydrogenase complex, complex II; ubiquinol:cytochrome c oxidoreductase, complex III; cytochrome c oxidase, complex IV; and the mitochondrial ATP synthase, mATP syn, complex V) and two mobile electron carriers (ubiquinone, UQ; cytochrome c, Cyt c) [319]. OXPHOS has rarely been studied in dinoflagellates. In the photosynthetic, Vitrella brassicaformis complex I is missing, while in Chromera velia and Amoebophrya ceratii this is the case for complexes I and III of the respiratory chain [253, 320]. Thus, one may speculate that also the respiratory chain of P. cordatum may feature a similar evolutionary adaptation, i.e., absence of a classical complex I. In accord, proteomic evidence for complexes II‒V was recently obtained for P. cordatum (Fig. 15b) [30].

The structure of chloroplasts and their embedded photosynthetic machineries have long been studied with the model plant A. thaliana [321] and cyanobacteria [322]. The photosynthetic machineries are typically composed of canonical PSI, PSII, cytb₆f complex, and the cATP syn. A recent study with P. cordatum [149] adds a new perspective in this context, as illustrated comparatively in Figure 16 for Synechococcus elongatus (cyanobacteria), C. reinhardtii (green algae), Phaeodactylum tricornutum (diatom), P. cordatum (dinoflagellate), and A. thaliana (land plant).

Cyanobacteria are an ecologically significant and morphologically diverse group of prokaryotes thriving in marine, freshwater, and terrestrial habitats [323, 324]. Marine members of the Synechococcus group are abundant cyanobacteria in the photosynthetic picoplankton in temperate and tropical regions of the oceans [325]. The freshwater model-cyanobacterium Synechococcus elongatus is ∼2 μm in length and has a rod-like cell shape (Fig. 16a) [326]. S. elongatus is a well-studied model for investigating photosynthesis, circadian clock, resuscitation from dormancy, and synthetic biology applications [327‒330].

Cyanobacteria have simpler thylakoid membrane structures as compared to those of chloroplasts, in that often two or three of these membranes completely line the inner face of the cyanobacterial cell envelope (Fig. 16b) [326]. These peripheral thylakoid membranes sometimes converge on the plasma membrane forming so-called thylapse structures [331, 332]. In the cytoplasm, cyanobacteria concentrate RuBisCO (together with CA) within carboxysomes, which are bacterial microcompartments enclosed by a proteinaceous, selectively permeable “membrane” (shell) [333]. Recent cryo-electron tomography of the α-carboxysomes from the marine cyanobacterium Cyanobium sp. PCC7001 revealed hexadecameric RuBisCO (L₈S₈) organized in three concentric layers [334], while it is structured as paracrystalline lattice in the β-carboxysomes of the freshwater cyanobacterium Synechococcus elongatus [335].

PSI and PSII do not co-localize in the thylakoid membrane (Fig. 16c) [336]. A PSI monomer consists of 10 subunits and is surrounded with numerous pigment-binding proteins each harboring a high number of chlorophylls and carotenoids; assembly to the most commonly occurring PSI trimer yields an estimated size of ∼1 MDa for the PSI-LHC complex [337, 338]. In the case of Thermosynechococcus vulcanus, a PSI monomer was shown to be associated with eight IsiA-chlorophyll-binding proteins (IsiA, iron-starvation-induced protein A) (Fig. 16d, upper panel) [339]. By contrast, the PSII is a homodimer, with each monomer consisting of 17 subunits. PSII bears fanned-out numerous pigment-binding proteins, which form large protein complexes (phycobilisomes) and consist of the three phycobili proteins allophycocyanin, phycocyanin, and phycoerythrin (Fig. 16d, lower panel) [324, 337, 340‒342]. Further stress proteins associated with the cyanobacterial photosynthetic machinery include early light-induced proteins and high light-inducible proteins [343, 344]. A recent observation concerning the symbiotic relationship of the N₂-fixing cyanobacterium Candidatus Atelocyanobacterium thalassa and its host, the haptophyte microalga Braarudosphaera bigelowii, revealed a fine-tuned metabolic and morphologic interrelationship to support synchronized growth [345, 346].

The green algae C. reinhardtii is encountered in a wide range of habitats and is often referred to as “green yeast” due to its rapid growth [347, 348]. C. reinhardtii is widely used as model organism to study oxygenic photosynthesis since green algae and land plants share a common ancestor [347, 348], and various other research topics such as cell cycle [349] and synthesis/turnover of lipid droplets [350]. A typical C. reinhardtii cell has a diameter of ∼10 μm, bears two polar flagella, and possesses a single, large, cup-shaped chloroplast, which is located in a posterior position (Fig. 16a) [347, 351, 352]. In the case of the green alga Micromonas sp., analysis of FIB/SEM-based image series revealed the chloroplast to occupy ∼30% of the cell volume [151].

In general, the chloroplast of green algae contains a single pyrenoid (Fig. 16b) and several starch granules [151, 347, 351]; notably, in Micromonas sp., the pyrenoid is completely enclosed by starch [151]. The thylakoid membranes of green algae are organized in grana-like stacks of variable length with no regular spacing (Fig. 16b) [353]. Contrasting the arrangement described above for cyanobacteria, the monomeric PSI is assumed to be located in the stromal thylakoid membranes, whereas the dimeric PSII is embedded in the thylakoid membranes of the grana stacks (Fig. 16c) [354, 355].

The monomeric PSI is composed of 14 subunits and surrounded by nine pigment-binding proteins harboring various pigments, including the chlorophylls a and b, β-carotene, loroxanthin, violaxanthin, and lutein (Fig. 16d, upper panel) [356]. By contrast, the dimeric PSII core is enclosed by heterotrimers (N-, M-, and S-trimers) of pigment-binding proteins harboring the pigments chlorophyll a and b, neoxanthin and lutein, and together forming the PSII-LHCII supercomplex (Fig. 16d, lower panel) [357‒359].

Diatoms form together with dinoflagellates the bulk of marine phytoplankton [35] and are globally distributed with prevalence in coastal regions [360]. Among them, the pennate diatom Phaeodactylum tricornutum is a polymorphic model-organism with variable cell shapes: benthic life-cycle stages predominantly correlate with oval cells, whereas for a pelagic (planktonic) lifestyle mostly fusiform or triradiate cell shapes are observed [360]. Irrespective of shape, a typical P. tricornutum cell is 10–20 µm in length [360] and bears a single, central chloroplast occupying ∼30% of the cell volume [151] (Fig. 16a). However, across diatoms the numbers, shapes and localizations of chloroplasts can vary profoundly [361]. In general, a special feature of diatom chloroplasts is the enclosure by four membranes; two of them are probably remnants of secondary endosymbiotic events [150, 151, 362].

In the majority of diatoms, three thylakoid membranes are arranged in stacks, which are interconnected forming an anastomosis [361], which yields lamellae triplets without grana stacks in the case of P. tricornutum (Fig. 16b) [362]. Pyrenoids for protein storage are submerged or semi-submerged into the chloroplast [361, 362].

Composition and spatial arrangement of the photosynthetic machinery of diatoms is essentially as described for cyanobacteria and green algae (Fig. 16c) [363, 364]. The monomeric PSI of the diatom Chaetoceros gracilis was recently found to be composed of 12 core subunits surrounded by 24 pigment-binding proteins with ∼500 associated pigments, including chlorophyll a and c₂, fucoxanthin, diadinoxanthin, and β-carotene [365, 366]. This PSI-LHCI supercomplex has an estimated size of ∼1.1 MDa (Fig. 16d; upper panel) [366]. By contrast, the dimeric PSII is composed of 19 subunits per monomer and enclosed by four hetero-tetrameric light-harvesting FCPs (S- and M-tetramer) and monomeric FCPs (FCP-DEF) (Fig. 16d; lower panel) [367].

A typical P. cordatum cell has a size of 10–20 µm (Fig. 16a) [116, 149]. 3D reconstruction of an entire P. cordatum cell based on FIB/SEM images revealed that a single, barrel-like chloroplast occupies ∼40% of the cell volume and completely lines the inner face of the cell’s envelope [149]. A similar chloroplast structure could be observed for the endosymbiotic dinoflagellate Symbiodinium, where the single chloroplast occupied ∼30% of the cell volume [151]. In general, the chloroplast envelope of dinoflagellates consists of three membranes as also known from Euglenophyta [113, 145, 368].

Similar to the chloroplast structure, also number and arrangement of the thylakoid membranes among dinoflagellates are highly diverse [145]. In general, several thylakoid membranes (2–4) form lamellae, which run parallel to the chloroplast without the formation of grana (Fig. 16b) [145]. Dinoflagellates contain pyrenoids for protein storage and harbor several starch inclusions [145].

The photosynthetic machinery of dinoflagellates is suggested to consist of similar complexes as aforementioned for the other organisms. However, in contrast to other known machineries, proteogenomic analysis of chloroplast-enriched fractions of P. cordatum revealed a light-harvesting-PSI/PSII megacomplex of ∼1.5 MDa estimated size (Fig. 12a, lower panel, 16c) [149].

The two photosystems of the megacomplex are presumed to be surrounded by numerous pigment-binding proteins equipped with a high number of diverse pigments including chlorophyll aand c₂, diadinoxanthin, violaxanthin, peridinin, and β-carotene (Fig. 16d) [149]. Noteworthy, the unicellular red alga Porphyridium purpureum, usually inhabiting moist terrestrial areas, also possesses a phycobilisome-PSII-PSI-LHC megacomplex, the structure of which was recently resolved at near-atomic resolution by cryo-electron microscopy [369].

Another example of a giant PSI/LHC supercomplex was recently reported for the endosymbiont Breviolum minutum (formerly Symbiodinium minutum), where the monomeric PSI is surrounded by 25 pigment-binding proteins harboring diverse pigments including chlorophyll a and c₂, diadinoxanthin, diatoxanthin, and peridinin [370]. Recent analysis of the PSI-AcpPCI supercomplex in Symbiodinium sp. GY-H50 by cryo-electron microscopy revealed that the tridecameric PSI core is surrounded by 13 peridinin-chla/c-binding light-harvesting antennae proteins (AcpPCIs) facilitating efficient energy transfer [371, 372].

As opposed to cyanobacteria and photosynthetic unicellular algae, land plants are multicellular organisms performing photosynthesis in the leaves. In the model plant A. thaliana, a mature leaf has a size of several cm and consists of around 500,000 cells (Fig. 16a) [373]. The outermost layer of the leaf is made up of an epidermis, a cuticle as well as a palisade mesophyll sub-layer, which is packed with cells [374, 375]. The palisade mesophyll harbors numerous chloroplasts, which are each enclosed by two membranes [374].

The thylakoid membranes of chloroplasts are partly organized in grana stacks (Fig. 16b) [374, 376]. Furthermore, chloroplasts of land plants also have starch inclusions but typically lack pyrenoids [377].

The composition of the photosynthetic machinery in land plants is in accord with the aforementioned ones plus the NADH-dehydrogenase-like complex type-1 (NDH-1; also termed chloroplast complex I, CCI) [378]. The PSI and PSII with their respective LHCs are usually spatially separated in the thylakoid membranes (Fig. 16c) [379]. While the monomeric PSI is embedded in the stromal thylakoid membranes, the dimeric PSII is located in the thylakoid membranes of the grana stacks [379]. However, a recent study showed that both photosystems apparently form PSI/PSII-supercomplexes at the margins of grana [380].

PSI and PSII each have a supramolecular organization [321]. The PSI predominantly occurs as monomer but can also form dimers or trimers [321]. Each monomer is surrounded by six pigment-binding proteins harboring chlorophylls (Fig. 16d, upper panel) [321, 378]. In the case of trimeric PSI, association with its LHC yields a supercomplex with an estimated size of ∼1.6 MDa [321]. The dimeric PSII is surrounded by four pigment (chlorophyll)-binding proteins forming a supercomplex of ∼1.3 MDa (Fig. 16d, lower panel) [321, 378].

The structure of chloroplasts among photosynthetic organisms is highly diverse and ranges from several small chloroplasts (e.g., land plants) to large, reticulate chloroplast networks (e.g., dinoflagellates presumably present under certain conditions). In addition, the increase in storage inclusions inside the chloroplast and a larger pyrenoid volume can change chloroplast structure even within the same species. These variations are not only attributable to morphological changes and adaptations of chloroplast structures during evolution, but also represent a response to various other factors such as lifestyle or prevailing environmental conditions [150, 151, 153]. The dinoflagellate Zooxanthella, e.g., forms larger chloroplast networks during endosymbiosis [153], and the green alga C. reinhardtii increases the volume of its pyrenoids in response to CO₂ limitation [156]. In parts, membranes of the chloroplast envelope are suggested to be remnants of secondary endosymbiosis events. The chloroplasts in diatoms are enclosed by four membranes, in dinoflagellates (also Euglenophyta) by three, and in plant cells by two. The comparison of the five selected photosynthetic model organisms (Fig. 16) illustrates distinct features of the photosynthetic machineries of limnic/marine microalgae compared to land plants, which might reflect, amongst others, adaptations to the markedly different habitats. This includes the organization of PSs as well as structure and composition of LHCs. Against this background, P. cordatum stands out in three ways: (i) unique, barrel-like structure of the chloroplast without grana, (ii) enormous range of pigment-binding proteins (140 predicted, 83 identified) providing high flexibility of light capturing, and (iii) unusual LHC-PSI/PSII megacomplex, which could improve photosynthetic efficiency [149].

Next to the ubiquitously occurring, bloom-forming P. cordatum several other model dinoflagellate species have been intensively investigated, suggesting themselves as promising model systems [381] and exemplified in the following for A. carterae, C. cohnii, members of the genus Symbiodinium, and Amoebophyra ceratii. The athecate (naked), marine, harmful bloom-forming A. carterae [132] has been widely studied, e.g., with respect to photosynthesis [372, 382], production of antifungal polyketides of the amphidinol family [383, 384], response to climate change conditions such as increasing temperature and acidification [385], as well as response to phosphorus deficiency [386], seasonality [387], and phycosphere [388].

The heterotrophic, non-photosynthetic C. cohnii [389] attracted early attention for its nuclear organization [390], including telomere maintenance [391], identification of a new class of transcription initiation factors [290], and cell cycle progression [212, 392]. C. cohnii stands out for its unusually high lipid content leading to continued biotechnological interest in the formation of the nutraceutically and pharmaceutically relevant docosahexaenoic acid [393, 394]. Furthermore, floridean starch synthesis from UDP-glucose, rather than via the ADP-glucose-based pathway used, e.g., by green algae, has been investigated [395].

Members of the genus Symbiodinium encompass free-living, opportunistic, and symbiotic species. The role of Symbiodinium sp. in symbiosis with reef-building corals prompted extensive genomic comparisons, which revealed adaptive evolution and divergence within the family Symbiodiniaceae and the genus Symbiodinium [24, 255, 396]. Immunological host determinants as well as metabolic exchange between Breviolum minutum (basionym: S. minutum) and the sea anemone Aiptasia pallida (synonym Exaiptasia) have been studied [397, 398]. In the case of Symbiodinium microadriaticum, the proteomic response to thermal and nutrient (nitrogen, phosphorus) stress has been catalogued [399]. Stress-resilience of the photo-physiological performance of S. microadriaticum is positively affected by the associated microbiome [400]. Aforementioned symbiotic members of the genus Symbiodinium harbor PSI-LHC supercomplexes [371, 372]. Further studies with B. minutum were concerned, e.g., with chromatin accessibility [401], differential gene expression in response to symbiotic lifestyle [402] and cell cycle [403], as well as establishment and/or maintenance of symbiosis independent of photosynthesis [404].

The deep-branching dinoflagellate Amoebophyra ceratii is an obligatory parasite of other dinoflagellates [405, 406]. An interesting observation reported by Kim and Park [407] is that Amoebophyra infections of the toxin-producing (shellfish poisoning) Alexandrium fundyense apparently reduces toxin production thereby promoting a healthy ecosystem. In the case of A. ceratii, functional mitochondria are apparently devoid of a mitochondrial genome, suggesting mitochondrial proteins to be encoded in the nucleus [253]. Comprehensive sequencing approaches showed that the Amoebophyra sp. complex related to Alexandrium minutum blooms harbors cryptic Amoebophyra species [408].

Dinoflagellates take over crucial functions in limnic and marine ecosystems on a global scale and stand out for their multifaceted morphology as well as special cell biology and metabolic capacities. The recent insights into P. cordatum summarized here open new avenues for future research centering around the intricate interplay between cellular dynamics and metabolic performance against the backdrop of adaptation to climate change conditions. Future research directions expanding the model character of P. cordatum may involve: (i) the nucleus with its still unclear process of dinomitosis and the poorly understood architectures of chromosomes and NPCs, (ii) the structural constancy versus variability of chloroplast and mitochondrion including shape and number, (iii) a detailed understanding of the composition and architecture of the light-harvesting PSI/PSII megacomplex and its diverse pigment-binding proteins, and (iv) the balance between photoautotrophic and mixotrophic metabolism. All of these topics need to consider the plethora of (mostly) external determinants, ranging from cell/diurnal cycles via organismic interactions (e.g., with heterotrophic pelagic bacteria, viruses, and general community dynamics), to changing environmental conditions (e.g., light regime, temperature, salinity, acidification, and nutrient availability). Building on the model P. cordatum, research should be expanded to other free-living dinoflagellates from freshwater as well as marine habitats. In a first step, extensive comparative genomics requires improved sequencing and assembly strategies to deliver the database for assessing fundamental communalities/differences with respect to genome/gene architecture, shaping by HGT, phylogenomics, etc. Furthermore, dissemination and variations concerning organelle structures and photosynthetic megacomplexes needs to the investigated. Ultimately, a reductionist knowledge base will be generated to advance our understanding on dinoflagellate activities and prospects in the environment.

We are grateful to M. Gottschling (Munich), M. Hoppenrath (Wilhelmshaven), E. Rhiel (Marburg), and M. Winklhofer (Oldenburg) for helpful comments on the manuscript.

The authors have no conflicts of interest to declare.

This manuscript (review) was written with no exteramural funding.

J.K. and R.R. conceived and wrote the manuscript.