Dinophysiales, known as dinophysoids, are among the most diverse group of dinoflagellates in terms of morphology. Their fine morphological gradients implicate evolutionary sequences, and are attractive to investigators. During the last few decades, the attention on this group rapidly increased, since some of the member species were found to produce toxins responsible for diarrheic shellfish poisoning.

The taxonomy of dinophysoids is, however, very problematic. Currently their classification is mainly based on cell size and shape, the presence of chloroplast, and cell ornamentations such as the presence and structure of lists, ribs, protrusion and spines. But, as the intraspecific variation is strong, these morphological characteristics are often overlapping among species. Species boundaries are thus ambiguous and specific assignment usually cannot be made with certainty.

A constraint for taxonomic study of these dinoflagellates is that most of them are rare and mainly distributed in pelagic waters. The sampling therefore faces many difficulties. Another problem is that the original diagnoses, which were mostly made 50 years or even more than a century ago, were usually very simple and not clear enough for specific separation. These difficulties hinder the understanding on the nature of species, making the classification system very artificial. Many taxonomists considered it is necessary to re-construct the classification of the order more natural. And DNA sequences are suggested to be a useful supporting tool.

This study aims to make the classification of the order clearer. By collecting material of a wide spectrum of taxa from various places and subjected to detailed morphological examination, including the plate patterns, and compared with the original description; and by analyzing their DNA sequences, I try to grasp the ranges of variations and figure out their specific boundaries, so as to make the classification of the order to be more robust.

The study area was in the pelagic waters of central and western Pacific Ocean collected during the two cruises of R/V MIRAI, MR07-01 and MR07-06, and some coastal waters of Japan, Thailand and Vietnam. Samples were collected by plankton net with a mesh size of 20 μm, either by towing net - in the case of coastal waters, or by filtering the seawater pumped up from vessel bottom while cruising - in the case of open ocean. Live cells of Dinophysiales were individually isolated under microscope, examined for their morphology, photographed and then transferred each to PCR tube, where they were thermally broken by several frozen-melt cycles, before directly subjected to DNA amplification. The rDNA of the D1-D2 domains of the large subunit were amplified by two rounds of PCR reaction. The products were purified and directly sequenced for both strands. The phylogenic trees were conferred by the neighbour-joining and the minimum evolutionary methods. Cells after analyzed for DNA were harvested for plate pattern examination. Through this process, the same cells were used for both the DNA analysis and the detailed morphological analysis.

A total of 57 species of Dinophysiales belonging to 9 genera, Dinophysis (25 species), Pseudophalacroma (2), Metaphalacroma (1), Metadinophysis (2), Ornithocercus (10), Citharistes (2), Histioneis (8), Amphisolenia (6), and Oxyphysis (1) were recorded, including many rare species. The existence and morphology of Dinophysis acutissima and, particularly, the genus Metadinophysis, were confirmed for the first time since their first description.

There are 3 morphotypes, Dinophysis sp., Histioneis sp. and Metadinophysis sp., do not fall into any previous descriptions, and thus must be established as new species. Each of the genera Metadinophysis and Pseudophalacroma consists of more than one none-conspecific morphotypes, indicating that these two genera are not monospecific.

Among the 38 species belonging to 9 genera examined for plate pattern, 20 species were observed for the first time. These observations resulted in some different conclusions from what previous taxonomists made.

(a) In all dinophysoids examined, the four sulcal plates were found to surround the flagellar pore. This is in contrast to conclusion made by previous study that the left sulcal plate of the genus Pseudophalacroma does not contact the flagellar pore, which was thought to be the distinct characteristic of the genus.

(b) Examination of plate pattern of Amphisolenia showed that the left sulcal plate in this genus was located right beside the flagellar pore, and never reached the far-away cingulum as it was assumed by previous authors. The two ventral hypothecal plates were arranged vertically, one after another, not horizontally as they were believed before.

(c) The genus Citharistes evidently had 4 cingular plates, not 6 plates as it was reported. This lead to an updated conclusion that all species of dinophysoids shared the similar plate pattern: 4 epithecal, 4 cingular, 4 sulcal and 4 hypothecal plates, in addition to an apical complex, which contained one or, usually, two pore platelets/plates.

Polymorphism - the formation of cells with different size and shape - was confirmed in 4 genera. Among them, genera Ornithocercus, Metadinophysis and Histioneis were for the first time observed to show this phenomenon. These recordings indicated that this phenomenon is very common in dinophysoids, and it was probably the main source of intraspecific variation in morphology of these dinoflagellates. The general rules of cell deformation during depauperating division were pointed out: cells always remained the anterior ventral part and discharged the dorsal posterior part. These understandings gave clues to guess the possible shape of small cell of certain species, once depauperating division undergoes.

The height of epitheca, the depth of cingulum, the deflection of body axis, thecal surface structure and the relative distance of rib 2 compared to rib 1 and rib 3 and the capability of hosting chloroplast were proved to be stable and specific for species, and they are therefore reliable criteria for classification. However, the judgment of these characteristics must be taken with full awareness of the effect of depauperating division on morphology of new cell, as mentioned above. In contrast, some characteristics, such as the structures of cingular list, the left sulcal list, were evidently not reliable characters, although in many cases they could provide hints for classification. This is because these characteristics could easily change according to the age of the cell and due to depauperating divisions.

Sequence of the D1-D2 domains turned out to be rather specific for species of Dinophysiales. The difference in DNA sequence usually correlated well with the morphological difference, indicating that this is a good criterion supporting the species differentiation in Dinophysiales.

Considering the above understandings, two trends of treatment on the classification of Dinophysiales were proposed, namely merging and splitting of taxa. Each of the species O. magnificus (with 3 morphotypes), D. fortii (2), D. infundibular (2) and D. hastata (5) , D. doryphora (2) contained several morphotypes, which were striking different in both morphology and genetic, and must be separated into new species. Merging was proposed to some other species that were differences in morphology, but those differences appeared to be resulted from polymorphism. H. hippoperoides, H. dolon and H. helenae should be considered as synonymous with Histioneis megalocopa, while O. steinii, O. skogsbergii, and probably also O. quadratus, may be conspecific with O. thumii; and various species of Dinophysis with minor differences in morphology, such as D. anabilis, D. lapidistrigiformis and D. microstrigiformis, may be conspecific with D. acuminata or D. fortii.

There had been morphological and molecular biological evidences for these treatments in the cases of D. infundibular, the D. hastata complex and the above mentioned Histioneis spp., but for other cases, further evidences are needed before such treatment could be made.

A well-known morphotype of Dinophysis, which is frequently referred as Dinophysis rotundata world wide, turned out to be not conspecific with the type specimen of D. rotundata Claparede et Lachmann, since they showed striking differences in thecal surface, cell shape and the height of epitheca, which are, as mentioned above, important specific characteristics. A new name therefore should be given to this morphotype. The same recommendation was raised for other two species, D. odiosa sensu Tai et Skogsberg, and D. elongata sensu Abe.

The study also revealed several important characteristics that were specific for genera or species, but were constantly overlooked or were not considered to be important by previous taxonomists. These included the position of the left ventral epithecal plate in Pseudophalacroma, the dissimilarity in convexity of the two main hypothecal plates of Metadinophysis, the deflection of body axis and the dorsal concavity of cingulum in Ornithocercus, and the epitheca depth in D. caudata-D. tripos complex.

The definition of O. quadratus was confined to morphotypes that corresponded to the type specimen only. Those were with different cell shape, which were mostly belonging to O. quadratus var. assimilalis, were excluded from this taxon and transferred to O. galea.

The three species, O. biclavatus, O. carolinae and O. galea, which had been repeatedly considered as conspecific with O. heteroporus, O. francescea and O. quadratus, respectively, were confirmed here to be valid independent species, since they showed difference in various important morphological criteria.

Most of the 79 sequences of 43 species of dinophysoids obtained in this study were for the first time to be read. Phylogenetic tree based on the obtained DNA sequence showed that the order Dinophysiales was monophylic. The combination of morphological and genetic evidences suggested that this group may have been evolved from some primitive forms like Phalacroma or some microcephalic group, such as Metaphalacroma or Metadinophysis etc. Among dinophysoids, the genus Amphisolenia, which were well defined in morphology and monophylic in genetic, appeared to diverge very early, in a separated direction. The rest of dinophysoids fell in 12 clades, the detailed arrangement of which was uncertain. The photosynthetic species of the genus Dinophysis sensu stricto evidently evolved independently. These species, which showed heavy thecal surface and had capability of holding chloroplast, formed a solid clade that was clearly separated from the rest of dinophysoids, including the rest of Dinophysis. Within this clade, the closely related groups D. caudata - D. tripos - D. miles formed their own evolutionary branch, with fine gradients in both genetic and morphology. Both morphological and genetic data supported the assumption by some previous taxonomists that Citharistes might have been evolved from the Ornithocercus stock, which might have been sprung from a certain form of Dinophysis, likely D. hastata. The participation of the genus Histioneis, which were likely monophylic in genetic, in this evolutionary sequence was not clear, despite their obvious intermediate morphology between Ornithocercus and Citharistes. Relationship of other groups, including the non-photosynthetic Dinophysis sensu stricto, the former members of Phalacroma and genus Oxyphysis, was uncertain.

In summary, with the results from examination of morphology and molecular biology of 57 dinophysoid species in Pacific region, the rules of morphological variation were understood, and the importance of some morphological and genetic characteristics for classification was realized. With these understandings, some confusion in classification of the order was clarified. Considering these approaches and understandings, by further study on other species of dinophysoids, as well as on materials from other areas, such as the Atlantic and the Indian Ocean regions etc., a more robust classification of the order can be obtained.

渦鞭毛藻は海洋の微小プランクトンの主要な一群であるだけでなく、赤潮や貝毒の原因となる有害種を多く含んでおり、多くの属種がさまざまな研究対象とされてきた。しかし、ディノフィシス目は、渦鞭毛藻綱の中でもっとも形態的変化にとみ、その連続性は進化との関係を想起させる一方、独立した種としての変異の限界が認めにくいため、研究対象としにくい分類群であった。しかもほとんどの種は外洋に分布しており、発生量が一般に極めて少なくいため、長い間分類学的研究すら行われていなかった。ただ、近年ディノフィシス目のDinophysis属の約十種に人に食中毒を引き起こす下痢性貝毒が発見され、しかも同一種の毒生産性に地域的差異がみられるため、分類や生態に強い関心が寄せられるようになって、改めてその分類学的研究の必要性が再認識されていた。このような背景のもと、本研究はディノフィシス目の多くの種の形態を精査して、有用な分類基準を確立した上で現在の分類体系を見直し、さらに分子生物学的形質を用いて形態に基づく分類の妥当性を確認しようとしたものである。

研究材料としては中部および西部太平洋の外洋域と日本・タイ・ベトナムの沿岸域で、孔径20マイクロメーターの目合いのプランクトンネットにより採集した試料を用いた。これら試料は、光学顕微鏡下で個体毎に細胞表面の鎧板配列と各鎧版の特徴を詳細に観察し、原記載など過去の形態観察例と比較して、種を査定した後、同一個体を用いてLSU rDNAのD1-D2領域とITS領域の塩基配列を決定した。目全体の系統解析は近隣結合法と最小進化法を用いて分析した。

本研究の結果、得られた試料から9属57種が確認され、その内訳は、Dinophysis 属25種、Pseudophalacroma属2種、Metaphalacroma属1種、Metadinophysis属2種、Ornithocercus属10種、Citharistes属2種、Histioneis属8種、Amphisolenia属6種、Oxyphysis属1種であった。これらには新種として記載するべき種や、他種の異名として整理すべき種がいくつかみられ、原記載後に今回初めて採集された希産種もあった。これらの分類の基準となる形質としては、上殻の高さ、横溝の深さ、体軸の傾き、殻板表面の模様構造、腹側第2肋と他の肋の間の相対距離、細胞内の色素体の特徴などが信頼できるものであった。観察された種のうち20種は鎧板配列が初めて詳細に調べられたもので、その結果、これまでに認められていたディノフィシス目の鎧板配列に関する解釈と異なる発見がいくつかあった。特にPseudophalacroma属のみ他属と異なるとされていた左縦溝板の位置、およびAmphisolenia属を他属と分けて独立した科にしていた、左縦溝板と2枚の腹側下殻板の位置について、従来の観察の誤りがあることを明確にすることができた。さらに、6枚の横溝板があるとされていたCitharistes属にも、従来の観察の誤りがあり、他属と同様横溝板は4枚であることを明らかにした。この結果、ディノフィシス目のすべての属種が、通常2枚の小板のある頂孔板群に加え、4上殻板、4横溝板、4縦溝板、4下殻板という相同の鎧板配列を共有していることが明らかになった。

また、同一種内で異なる大きさと形態の細胞が形成される生物学的多型現象が目内の4属で確認され、この現象がディノフィシス目で多くの種に共通な特徴である可能性が示された。さらに43種で79のD1-D2領域の塩基配列が認められたが、この解析からディノフィシス目が単系統であることが示された。また、この配列における差異の有無は形態の差異の有無とよく関連しており、種の分化を支持する良い判断基準であるだけでなく、その系統解析結果は形態により推定される種の分化とよく一致することが分かった。

以上本研究は、その困難さから研究の立ち遅れていたディノフィシス目の渦鞭毛藻類に関し、沿岸と外洋域で約60種の詳細な形態観察を基に重要な分類形質を定め、その基準に基づきディノフィシス目の分類体系を検討し、さらに分子生物学的手法によりその分類体系の妥当性を検討したものである。この成果により、ディノフィシス目渦鞭毛藻類の分類の基礎が築かれ、今後の世界各地で採集される有毒種を含めたディノフィシス目渦鞭毛藻の生理生態、生化学研究において研究材料の同定の標準化を諮ることが容易になり、水圏生物科学の発展に資するところが極めて大であると考えられた。よって、審査委員一同は本論文をもって、本研究の申請者ニュエン バン ニュエン君が博士(農学)の学位を授与するに値する研究者であると判断した。