This account firstly outlines the relationships between algal diversity and chloroplast acquisition through endosymbiosis. Secondly, it briefly reviews chloroplast diversity in dinoflagellates. Particular emphasis is placed on the evolutionary process in the small but interesting group of dinoflagellates that possess a diatom endosymbiont.
Diversity of eukaryotic algae
The eukaryotic algae are an assemblage of chloroplast-bearing photosynthetic organisms. They are extremely diverse in terms of morphology, cytology and life cycle (e.g., Hoek et al., 1995). The eukaryotic algae can be classified into nine divisions, the Rhodophyta, Chlorophyta, Glaucophyta, Heterokontophyta (including Phaeophyceae (brown algae), Bacillariophyceae (diatoms), Chrysophyceae, Raphidophyceae, Xanthophyceae and other minor groups), Haptophyta, Cryptophyta, Dinophyta, Euglenophyta and Chlorarachniophyta. Recent molecular phylogenies indicate that, although each group is monophyletic, many of them are totally unrelated, with closer relationships to nonphotosynthetic protists (e.g., Euglenophyta and Kinetoplastida; Dinophyta and Ciliophora; etc.) (e.g., Cavalier-Smith, 1993a). The presence of chloroplasts in these unrelated groups is due to chloroplast acquisition through independent endosymbioses.
The phylogenetic affinities of eukaryotic organisms, as a whole, are becoming clearer by the advancement of molecular phylogenetic studies, although discrepancies between phylogenies derived from different genes are encountered and some uncertainty exists with regard to the order of deep branches. Recent schemes for the classification of the eukaroyotes recognize six “kingdoms” (Simpson and Roger, 2004a), viz., Opisthokonta (e.g., animals, choanoflagellates and fungi), Amoebozoa (e.g., slime molds and lobose amoebae), Plantae (e.g., land plants and green algae, red algae and glaucophytes), Chromalveolata (e.g., Chromista [Hetrokontophyta, Haptophyta, Cryptophyta, oomycetes, bicosoecids and labyrinthulids] and Alveolata [dinoflagellates, apicomplexa and ciliates]), Rhizaria (e.g., Cercozoa and Radiolaria) and Excavata (e.g., Euglenoza, jacobids and diplomonads) or five “supergroups” (Keeling, 2004), viz., Unikonts (including Opisthokonta and Amoebozoa), Plantae, Chromalveolates, Rhizaria and Excavates (also see Cavalier-Smith, 2003, and Nozaki, 2005, for higher taxonomic systems of eukaryotes). Not all the groups described above, however, enjoy full support from the molecular data—the Excavata, for example, are recognized only when the molecular and morphological data are combined (Simpson and Roger, 2004b). By positioning the organisms termed ‘algae’ on this global phylogenetic tree of the eukaryotes, it readily become apparent how diverse the algae are (for example, see Keeling, 2004, Figure 1).
Figure 1.
A hypothetical evolutionary sequence of acquisition of secondary chloroplast (Chromista-type) through secondary endosymbiosis (based on Cavalier-Smith, 1993b). A photosynthetic alga with primary plastid is engulfed by a heterotrophic eukaryote and retained in a food vacuole of the latter organisms (A and B). The food vacuole (phagosomal membrane) is fused with the outer nuclear envelope and the organelles of the engulfed alga disappear (C) (at this stage, the vestigial nucleus (nucleomorph) is retained). Finally, the secondary chloroplast surrounded by four membranes is established (D). HN: host nucleus, EN: endosymbiont nucleus.
![i1342-8144-10-4-299-f01.gif](ContentImages/Journals/jpal/10/4/prpsj.10.299/graphic/WebImages/i1342-8144-10-4-299-f01.gif)
Origin of the chloroplast
The obvious common feature that is shared by eukaryotic algae is the presence of chloroplasts. The mechanism and timing of chloroplast acquisition in each group are the key issues for understanding of the origins of algae. It is widely accepted that the chloroplast was acquired originally through endosymbiosis, i.e., a primitive heterotrophic eukaryote engulfed a cyanobacterium-like photosynthetic prokaryote and eventually converted the endosymbiont into a photosynthetic organelle (see Tomitani in this volume and references therein). This process, resulting in the establishment of the first plant (alga), probably took place only once in the history of life (see Archibald and Keeling, 2004; Keeling, 2004). This process is called chloroplast acquisition via “primary endosymbiosis” and chloroplasts derived thus are called “primary chloroplasts.” The molecular data suggest that only three extant photosynthetic groups are direct descendents of this first plant. These include the Chlorophyta (including land plants), the Rhodophyta and the Glaucophyta. Because these three groups are very different in terms of morphology, photosynthetic pigment composition and mode of life cycles, e.g., the red algae (Rhodophyta) are totally devoid of flagellated stages, it is hard to believe that these three groups originate from a single ancestor. Such a monophyletic relationship is supported by molecular work (Rodoriguez-Ezpeleta et al., 2005), although Nozaki (2005) proposed that these three groups were paraphyletic, situating them in a basal position of his kingdom “Plantae.”
Origins of other algal chloroplasts
It is also widely accepted that chloroplasts in algae other than the three groups described above were established by secondary endosymbioses (Archibald and Keeling, 2004; Keeling, 2004). In secondary endosymbiosis, a phototrophic eukaryotic alga (with primary chloroplasts) was engulfed by a heterotrophic eukaryotic host and retained as an endosymbiont. Subsequently, the organelles of the endosymbiont (excepting the chloroplasts) were eventually lost (Cavalier-Smith, 1993b) (Figure 1). In the case of secondary endosymbioses, the resulting chloroplasts, i.e., secondary chloroplasts, are enclosed by either three (Euglenophyta and Dinophyta) or four membranes (Heterokontophyta, Haptophyta, Cryptophyta and Chlorarachiniophyta), while in primary chloroplasts (Chlorophyta, Rhodophyta and Glaucophyta) they are bounded by only two membranes. Two algal groups with green chloroplasts, the Euglenophyta and the Chlorarachniophyta, are thought to have acquired their chloroplasts from green algae, while some other groups, the Heterokontophyta, Haptophyta, Cryptophyta, and Dinophyta, are thought to have obtained their chloroplasts from red algae. How many times secondary endosymbioses have taken place for green and red lineages, respectively, is still under discussion (e.g., Cavalier-Smith, 2003, Bachvaroff et al., 2005).
Members of the Cryptophyta and the Chlorarachniophyta are interesting as they retain, in addition to chloroplasts, a reduced nucleus (nucleomorph) of the endosymbiont, exhibiting an intermediate evolutionary stage. The genomes of these two groups are well studied (Douglas et al., 2001; Gilson and McFadden, 2002). Although not usually regarded as algae, members of the Apicomplexa, a group of parasitic organisms including malarian parasites, are known to possess reduced chloroplasts, called apicoplasts (McFadden et al., 1996; Maréchal and Cesbron-Delauw, 2001). The apicoplast is also regarded as having been obtained via secondary endosymbiosis (Fast et al., 2001).
Recently, a number of reviews dealing with the origins of secondary chloroplasts, the number of secondary endosymbioses and the validity of the “Chromalveolata hypothesis” (Cavaler-Smith, 2004) have been published (for details, see Stobe and Maier (2002), Palmer (2003), Archibald and Keeling (2004), Keeling (2004), Bachwaroff et al. (2005), Harper et al. (2005), Nozaki (2005) and Yoon et al. (2005)).
For the successful acquisition of secondary chloroplasts, a protein targeting system must be established and this topic is reviewed by Ishida (2005).
Of all photosynthetic eukaryotes, the dinoflagellates are probably the most interesting with respect to chloroplast evolution because of their great diversity of chloroplast type. They therefore are briefly reviewed here.
Chloroplast diversity in dinoflagellates
The dinoflagellates are important members of aquatic ecosystems and some of them are known to produce red tides (Taylor and Pollingher, 1987). The dinoflagellates also produce cysts that preserve well and which have thus been well studied from a paleontological point of view (Fensome et al., 1993).
Different groups of dinoflagellates can have quite different types of chloroplasts. Obviously, the diversity of chloroplasts in dinoflagellates is due to the acquisition of different types of photosynthetic endosymbionts (Larsen, 1992; Schnepf, 1993; Schnepf and Elbrächter, 1999; Stoebe and Maier, 2002; Archibald and Keeling, 2004; Hackett et al., 2004). It should also be noted that only roughly half of known living dinoflagellate species actually possess chloroplasts; the rest exhibit heterotrophic nutrition. Additionally, plastid losses have been suggested to have occurred independently multiple times in dinoflagellates (Saldarriaga et al., 2001).
The types of chloroplasts found in dinoflagellates are classified into five categories: 1) a typical dinoflagellate “peridinin-type” chloroplast, 2) chloroplasts of green algal origin (Figure 2), 3) chloroplasts of haptophyte origin (Figure 3), 4) chloroplasts of diatom origin (Figure 4), and 5) kleptochloroplasts (Figure 5) and foreign algal components, whose level of permanency in incorporation into the host dinoflagellate (i.e., organelle versus food) is still uncertain.
Figures 2–5.
Dinoflagellates with unusual “chloroplasts”.
Figure 2. Lepidodinium viride with green chloroplasts.
Figure 3. Karenia mikimotoi with haptophyte-type chloroplasts.
Figure 4. Peridinium quinquecorne with a diatom endosymbiont. Note the prominent eyespot (arrow).
Figure 5. Amphidinium latum, fed with cryptomonad cells (kleptochloroplasts). This particular cell contains more than 10 cryptomonad cells. The organism can be maintained only by feeding with cryptomonad cells. Scale bars = 10 μm.
![i1342-8144-10-4-299-f02.jpg](ContentImages/Journals/jpal/10/4/prpsj.10.299/graphic/WebImages/i1342-8144-10-4-299-f02.jpg)
Peridinin-type chloroplasts
This is a typical chloroplast type for the dinoflagellates and the large majority of dinoflagellates possess this type of plastid. Even “typical” dinoflagellate chloroplasts are unique compared with those of other algal groups. The chloroplast is enveloped by three membranes and typically each lamella consists of three appressed thylakoids (Dodge, 1971; Schnepf and Elbrächter, 1999). The chloroplast contains, in addition to chlorophylls a and c, a xanthophyll, peridinin, which is unique to the dinoflagellates among extant organisms (Prézelin, 1987). The peridinin chloroplasts contain uniquely organized, circular plastid genomes, called “minicircles” (Zhang et al., 1999, Green, 2004). Dinoflagellates with peridinin-type chloroplasts possess Type II RUBISCO, instead of the usual Type I, and the gene for this protein is encoded in the nucleus rather than in the plastid genome (Morse et al., 1995). Peridinin chloroplasts are thought to be derived from a red alga through a secondary endosymbiosis (Zhang et al., 1999; Takishita and Uchida, 1999; Ishida and Green, 2002). For information on the operational targeting system for proteins to chloroplasts in dinoflagellates, see Patron et al. (2005).
Replacement of peridinin-type chloroplasts by those of newly acquired endosymbionts
The acquisition of chloroplasts by engulfing an alga with secondary chloroplasts (e.g., a haptophyte or diatom) is often called “tertiary endosymbiosis”. The phenomenon of tertiary endosymbiosis is only known to occur in dinoflagellates. An interesting evolutionary condition relating to these unusual (non-peridinin-type) chloroplasts is that their acquisition took place in lineages which already had peridinin-type chloroplasts (Saldarriaga et al., 2001, Taylor, 2004). In dinoflagellates, it is not unusual for chloroplast-bearing species to exhibit feeding behaviour (e.g., Wilcox and Wedemayer, 1991) and this is obviously a prerequisite for the establishment of tertiary endosymbiotic chloroplasts.
Chloroplasts of green algal origin: Lepidodinium viride Watanabe et al. (Watanabe et al., 1990) (Figure 2) and Gymnodinium chlorophorum Elbrächter et Schnepf (Elbrächter and Schnepf, 1996) have green chloroplasts. The chloroplasts of these dinoflagellates are most likely to be derived from a prasinophyte alga (Elbrächter and Schnepf, 1996; Schnepf and Elbrächter, 1999; Taylor, 2004). Molecular analyses indicate that these two species are included in a Gymnodinium clade (Daugbjerg et al., 2000; Salddariaga et al., 2001). Morphologically these two species exhibit features of the genus Gymnodinium sensu Daugbjerg et al. (2000), although ultrastructural studies indicate that they are somewhat different from the type species of the genus Gymnodinium, G. fuscum (Watanabe et al., 1990; Hansen and Mestrup, 2005). It should be noted that the acquisition of green plastids does not represent a tertiary endosymbiosis—it is actually a successive secondary endosymbiosis.
Chloroplasts of haptophyte origin: This type of chloroplast has been found mainly in the planktonic marine dinoflagellates (Figure 3). These dinoflagellates are characterized by the possession of chloroplasts with the photosynthetic pigments fucoxanthin, 19′-hexanoyloxy-fucoxanthin and/or, 19′-butanoyloxy-fucoxanthin and have no peridinin (Daugbjerg et al., 2000, de Salas et al., 2003). No other organelles thought to be derived from the endosymbiont have been detected in the dinoflagellate cytoplasm. Recently, the amount of information on these dinoflagellates has increased. More and more new taxa have been described during the last few years and consequently three genera (Karenia G. Hansen et Moestrup, Karlodinium J. Larsen and Takayama de Salas, Bolch, Botes et Hallegraeff), comprising almost 20 species, are recognized (see Gómez, 2005). Since the species of these three genera form a single clade, it is quite obvious that they evolved from a single ancestor (de Salas et al., 2003). Analyses of molecular data regarding the evolutionary origin of this unusual chloroplast type have been published (Tengs et al., 2000; Ishida and Green, 2002; Yoon et al., 2005). Ishida and Green (2002) demonstrated that the dinoflagellate nuclear psbO (oxygen-evolving enhancer 1) gene of Karenia brevis has been replaced by an analogue of a haptophyte endosymbiont. According to recent reports, it is almost certain that the origin of this type of chloroplast is a haptophyte.
Kleptochloroplasts or true tertiary endosymbioses?: In addition to dinoflagellates with permanent chloroplasts (or permanent endosymbioses) there are dinoflagellates which make use of “temporary chloroplasts.” Temporary chloroplasts are also called “kleptochloroplasts (kleptoplastids) = stolen chloroplasts” (Schnepf, 1993) and several different taxonomic groups of dinoflagellates are known to exhibit this phenomenon (Figure 5) (Larsen, 1988; Schnepf et al., 1989; Fields and Rhodes, 1991; Horiguchi and Pienaar, 1992; Takishita et al., 2002; Lewitus et al., 1999; Koike et al., 2005).
Some species of Dinophysis are known to harbour cryptomonad-like chloroplasts with a double envelope (Schnepf and Elbrächter, 1988). Takishita et al. (2002) discovered that the SSU rDNA sequences from the plastids of three species of Dinophysis are identical, but that there are some substitutions in their nuclear SSU rDNA. It is impossible to grow members of this genus in culture and this, together with the unexpected finding of identical plastidial genomes in different species, led Takishita et al. (2002) to suggest that the chloroplasts in Dinophysis are probably kleptochloroplasts and not true chloroplasts. Recently, Koike et al. (2005) reported the presence of haptophyte algal plastids in Dinophysis mitra and this relationship is another example of a kleptochloroplast. Schweikert and Elbrächter (2004) reported the presence of algal cellular components in Podolampas bipes. Each dinoflagellate cell contains some 10 foreign algal cells. Ultrastructural studies indicate that the incorporated cells possess almost all of their cellular components and have affinity with members of the Pedinellales. Whether these foreign cells represent true endosymbiotic partners still needs to be confirmed.
Diversity of dinoflagellates with a diatom endosymbiont: The following eight species are now recognized as dinoflagellates with a diatom endosymbiont: Durinskia baltica (Levander) Carty et Cox (= Peridinium balticum (Levander) Lemmermann) (Tomas and Cox, 1973), Kryptoperidinium foliaceum (Jeffrey and Vesk, 1976, Dodge, 1983, Kempton et al., 2002), Durinskia sp. (Pienaar, 1980; Horiguchi and Pienaar, 1988), Peridinium quinquecorne Abé (Horiguchi and Pienaar, 1991) (Figure 4), Dinothrix paradoxa Pascher (Horiguchi, 1993), Gymnodinium quadrilobatum Horiguchi et Pienaar (Horiguchi and Pienaar, 1994), Galeidinium rugatum Tamura et Horiguchi (Tamura et al., 2005) and a freshwater species belonging to the genus Peridiniopsis, Peridiniopsis sp. 1 (Takano et al., unpublished data).
Except for the first two, well known species, all have been demonstrated to possess an endosymbiont of diatom origin by Horiguchi and his coworkers. Although successful culturing is limited only to K. foliaceum, D. baltica and G. rugatum, these endosymbiotic associations are thought to be permanent and not the product of kleptochloroplasts. All the dinoflagellates in this group share the following characteristics (Figure 6): 1) the endosymbiont cytoplasm is separated from the dinoflagellate cytoplasm by a single unit membrane, 2) the endosymbiont cytoplasm contains chloroplasts, a nucleus, mitochondria and ribosomes, and 3) the host dinoflagellate cell possesses a particular type of eyespot (Type B eyespot, see below), located in the cytoplasm. In contrast to dinoflagellates with haptophyte chloroplasts, which are relatively uniform in morphology and in mode of life cycles, these eight species display an impressive variation. In fact, using present-day taxonomy, some species would be classified in different orders. In order to demonstrate how diverse these dinoflagellates are, the morphology and type of life cycle of these species will be briefly described.
Figure 6.
A schematic drawing to show the cellular structure of a dinoflagellate with an endosymbiont of diatom origin. The endosymbiont is separated from the dinoflagellate cytoplasm by a unit membrane (M). The endosymbiont contains a nucleus (EN), mitochondria (Em) and chloroplasts (C). The chloroplast ER and the connection between the former and the outer nuclear envelope is not illustrated. The dinoflagellate cytoplasm contains a nucleus (Dn), Golgi body (G), eyespot (E), trichocysts (T) and starch grains (s).
![i1342-8144-10-4-299-f06.gif](ContentImages/Journals/jpal/10/4/prpsj.10.299/graphic/WebImages/i1342-8144-10-4-299-f06.gif)
Durinskia baltica (Levander) Carty et Cox (order Peridiniales) (Figure 8E)
This small, motile dinoflagellate belonging to the order Peridiniales is the type species of the genus and has a thecal plate arrangement of: Po, x, 4′, 2a, 6″, 5c, 4s, 5′″, 2″″. It is found in both marine (Okinawa, Japan) and freshwater (Hokkaido, Japan) environments. This is a well known diatom-harbouring dinoflagellate species which has been extensively studied.
Durinskia sp. (order Peridiniales) (Figure 8F)
This species has been found in tide pools in several localities in South Africa, especially along the Cape coast. Its thecal plate arrangement is: Po, x, 4′, 2a, 6″, 5c, 5s, 5′″, 2″″ and, although this arrangement is almost the same as that of the type species, the gross morphology is very different, making it likely that it is a new species. The formal description of this species is now in preparation.
Kryptoperidinium foliaceum (Stein) Lindemann (order Peridiniales) (Figure 8B)
A highly flattened, motile marine dinoflagellate, which is also a well known and extensively studied diatom-harbouring species. The thecal plate arrangement is: Po, x, 4′, 2a, 7″, 5c, 5s, 5′″, 2″″, which is quite similar to that of Durinskia. However, the number of precingular plates is different (7 versus 6).
Peridinium quinquecorne Abé (order Peridiniales) (Figure 8D)
A characteristic motile dinoflagellate with prominent antapical spines and a thecal plate arrangement of: Po, x, 3′, 2a, 7″, 5c, 4s, 5′″, 2″″. This species has been found in numerous localities along the Japanese coast, including brackish Lake Hamana, Shizuoka Prefecture.
Peridiniopsis sp. 1 (order Peridiniales) (Figure 8G)
This species was discovered producing blooms in a freshwater pond in Toyama, central Japan in the spring/summer of the year 2003. The thecal plate arrangement is: Po, x, 4′, 0a, 6″, ?c. ?s, 5′″, 2″″. A detailed study of this species is in progress.
Gymnodinium quadrilobatum Horiguchi et Pienaar (order Phytodiniales?) (Figure 8H)
This species was isolated from beach sand sampled from the subtropical Indian Ocean coast of South Africa, but has subsequently also been found in a mud sample from a mangrove river on Iriomote Island, Okinawa, Japan. It is characterized by a dominant nonmotile cell with a four-leaf-clover-like gross morphology. This species reproduces asexually by forming two motile cells. The motile cell is athecate and Gymnodinium-like in morphology. Horiguchi and Pienaar (1994) described it as a new species of the genus Gymnodinium based on the motile cell morphology, rather than creating a new genus based on its unique vegetative morphology. However, now that the genus Gymnodinium is strictly defined (Daugbjerg et al., 2000), the motile cell of this species is clearly no longer a member. This species is here tentatively treated as a member of the order Phytodiniales, which is characterized by having a dominant nonmotile vegetative stage. Unfortunately, we have not been able to sequence this species to date.
Galeidinium rugatum Tamura et Horiguchi (order Phytodiniales) (Figure 8C)
This newly isolated strain from the sandy bottom of Jellyfish Lake, Palau, possesses a very unique, helmet-shaped vegetative form. The dominant stage in the life cycle is nonmotile and it firmly attaches to the substratum. It reproduces by the production of two motile cells, which are athecate. The motile form is short-lived, almost directly returning to the nonmotile vegetative form. This species has been recently described formally (Tamura et al., 2005).
Dinothrix paradoxa Pascher (order Dinotrichales, or family Dinocloniaceae, order Phytodiniales) (Figure 8A)
This dinoflagellate is characterized by its pseudofilamentous form, which consists of up to 10 spherical cells, formed by successive divisions of nonmotile cells. This is remarkable as, in the other dinoflagellates mentioned above, reproduction is always accompanied by the formation of motile cells. This dinoflagellate does, however, also have the ability to produce thecate, motile cells.
Evolution of dinoflagellates with a diatom endosymbiont: With reference to the variation in morphology and mode of life cycles at least some species of dinoflagellate with a diatom endosymbiont seem to be distantly related to each other. Conversely, a number of shared cytological features seem to suggest their common ancestry. This makes it clear that the possible monophyly of these eight species should be investigated using molecular data. Of the eight species described above, only two (D. baltica and K. foliaceum) already have been analysed using molecular data (Chesnick et al., 1996, 1997; Saldarriaga et al., 2001;Inagaki et al., 2000). However, molecular data for the other species recently have become available (Tamura et al., 2005). Apart from phylogenetic information, the molecular genetics of the endosymbionts of these dinoflagellates remain poorly studied relative to the nucleomorphs of cryptophytes or chlorarachiniophytes. Recently, however, McEwan and Keeling (2004) have demonstrated that actin, α-tubulin and HSP90 genes are present in the endosymbiont nucleus. This line of study will reveal, at a molecular level, what kind of changes have taken place during the process of tertiary endosymbiosis.
In discussing the origin(s) and evolutionary process(es) of these eight dinoflagellates species, some consideration of the structure called the eyespot would be useful. The eyespot is a structure related to phototaxis and it is widely distributed in motile cells of many groups of algae (Kawai and Kreimer, 2000). It is interesting to point out that all these eight species and only these species in dinoflagellates (as far as we are aware) possess the same type of eyespot. In this eye-spot (hereafter referred to as a “Type B eyespot, Dodge, 1984) rows of red-pigmented globules are bounded by a triple membrane and this is quite different from eyespots of typical dinoflagellates. The typical dinoflagellate eyespot (Type C eyespot, Dodge, 1984) is a row (or rows) of red-pigmented lipid globules located within the chloroplast, i.e., the eye-spot forms part of the chloroplast. It should be noted that the chloroplast of typical dinoflagellates is also bounded by a triple membrane. Because both Type B eyespots and typical dinoflagellate chloroplasts are bounded by triple membranes, it has been assumed that these two structures are homologous (Dodge, 1987). The following hypothesis to explain the origin of dinoflagellates possessing a diatom endosymbiont (and the origin of the Type B eyespot) has been proposed (Cavalier-Smith, 1993b) (Figure 7). Originally the ancestral dinoflagellate possessed a typical peridinin-containing dinoflagellate chloroplast with a Type C eyespot. This dinoflagellate subsequently engulfed a diatom and kept it as an endosymbiont. Molecular studies on diatom endosymbionts reveal that the diatom engulfed by the ancestral dinoflagellate was probably a benthic pennate diatom of the family Bacillariaceae (Chesnik et al., 1997), most probably a member of the genus Nitzschia (McEwan and Keeling, 2004; our unpublished data) (but see below). The original dinoflagellate chloroplast (peridinin-type) was subsequently mostly replaced by the diatom endosymbiont, leaving only the eyespot part of the original chloroplast as a device for phototaxis. The Type B eyespot is, therefore, a highly reduced typical dinoflagellate chloroplast (Dodge, 1987; Cavalier-Smith, 1993b). If this hypothesis is true, and the author believes it is, it is hard to accept that this kind of complicated evolutionary event took place several times independently. Rather, it is much simpler (and parsimonious) to think that the above-mentioned evolutionary process took place only once and later species with a diatom endosymbiont diversified. Horiguchi and Pienaar (1994) first proposed a single origin for all diatom-harbouring dinoflagellates, but at that time did not have the means to prove it. To test this hypothesis, molecular phylogenetic approaches are appropriate. A close affinity of D. baltica and K. foliaceum already has been demonstrated by Inagaki et al. (2000) based on SSU rDNA phylogeny. Tamura et al. (2005), also using SSU rDNA sequences, demonstrated that their new dinoflagellate Galeidinium rugatum forms a clade with K. foliaceum and D. baltica. The life cycle of G. rugatum consists of a dominant benthic, dome-shaped vegetative stage (Figure 8C) and naked (athecate), short-lasting motile cells, which directly return to a nonmotile, vegetative form (Tamura et al., 2005). This newly found taxon is, therefore, very different from K. foliaceum and D. baltica, which have more or less “normal” dinoflagellate features (apart from the presence of an endosymbiont). Interestingly, this benthic, crustose dinoflagellate is shown to be more closely related to the motile K. foliaceum than the motile D. baltica is to K. foliaceum. These results indicate that, regardless of form and mode of life cycles, dinoflagellates with a diatom endosymbiont are indeed monophyletic. We are now sequencing further nuclear-encoded SSU rDNA of other dinoflagellates with a diatom endosymbiont (except for G. quadrilobatum). Although the analysis is preliminary, our tree indicates that other diatom-bearing dinoflagellates also form a clade with K. foliaceum, D. baltica and G. rugatum (the results will be published elsewhere) (Figure 8). It is therefore probable, as proposed by Horiguchi and Pienaar (1994), that all diatom-harbouring dinoflagellates have evolved from a single ancestor after acquiring the diatom endosymbiont. Subsequently the species diversified, sometimes with radical changes in morphology (e.g., development of different thecal plate patterns or loss of thecal plates) or in mode of life cycles (e.g., adaptation to benthic life style). It should be noted that these characteristics (presence or absence of thecal plates and mode of life cycles) often have been used in dinoflagellate taxonomy at higher ranks, e.g., the order. Our series of works demonstrated that these changes were feasible within a single lineage and therefore that the validity of these characteristics as higher taxonomic criteria should be reconsidered.
Figure 7.
A hypothetical evolutionary process of establishment of a dinoflagellate with a diatom endosymbiont. The original dinoflagellate (A) probably possessed a typical peridinin-type chloroplast with Type C eyespot. A pennate benthic diatom (Nitzschia?) was engulfed by the dinoflagellate. The original peridinin-type chloroplast was replaced by the diatom endosymbiont (B) and subsequently all the photosynthetic functions were taken over by the endosymbiont, and only a part of the eyespot of the original chloroplast was retained (establishment of Type B eyespot: C).
![i1342-8144-10-4-299-f07.gif](ContentImages/Journals/jpal/10/4/prpsj.10.299/graphic/WebImages/i1342-8144-10-4-299-f07.gif)
Figure 8.
Preliminary phylogenetic tree of 7 out of 8 diatom endosymbiont species. The tree is based on our unpublished SSU rDNA analyses. Each vertical bar indicates: 1. gain of benthic life form, 2. loss of thecal plates and 3. gain of ability to divide continuously in nonmotile stage.
![i1342-8144-10-4-299-f08.gif](ContentImages/Journals/jpal/10/4/prpsj.10.299/graphic/WebImages/i1342-8144-10-4-299-f08.gif)
Another surprise: In addition to nuclear-encoded SSU rDNA, we have sequenced SSU rDNA of the diatom endosymbiont in Peridinium quinqecorne and the results of phylogenetic analyses were rather unexpected (Horiguchi and Takano, 2006). Our phylogenetic trees indicated that this endosymbiont did not form a clade with the endosymbionts of other diatom-bearing dinoflagellates (D. baltica and K. foliaceum), which clearly show affinity to members of the pennate diatom genus Nitzschia. Rather, it formed a strong clade with members of the centric diatom genus Chaetoceros (Horiguchi and Takano, 2006). The non-monophyly of endosymbionts seems to contradict the above hypothesis of a single endoymbiotic event. However, it should be emphasized again that the nuclear-encoded (host) SSU rDNA indicated that all the dinoflagellates with a diatom endosymbiont, including P. quinquecorne, are monophyletic. The only logical explanation for this is that, in P. quinquecorne, the original Nitzschia-like endosymbiont was later replaced by a member of Chaetoceros. The chloroplasts of dinoflagellates are, in many ways, full of surprises and this serial replacement of diatom endosymbiont is surely another surprising event.
Acknowledgments
I would like to thank Hiroshi Kitazato (JAMSTEC) and Hiroshi Nishi (Hokkaido University) for giving me the opportunity to present this paper at the symposium of the Paleontological Society of Japan. Thanks are due to Stuart D. Sym for kindly reading the manuscript. This work was partly supported by the 21st Century Center of Excellence (COE) Program on “Neo-Science of Natural History” (Program Leader: Hisatake Okada) at Hokkaido University financed by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. The work was also financed by Grants-in-Aid (12NP0102, 16370039) from the MEXT.
References
Appendices
Note added in proofs
The species described in this paper as Durinski sp. has now been described formally as Durinskia capensis Pienaar, Sakai et Horiguchi (2006).