Evolution, species resolution and molecular population genetics of the Gymnodinium catenatum toxic dinoflagellate species complex : tracing global dispersal and population dynamics

The chain-forming toxic dinoflagellate Gymnodinium catenatum Graham is a known causative organism of paralytic shellfish poisoning (PSP). During the mid-1970's the geographic extent of G. catenatum plankton blooms increased dramatically, causing toxic episodes in Spain, Portugal, Mexico, Venezuela, Argentina, Uruguay, Japan, Korea and southern Australia (Tasmania). Studies of the distinctive microreticulate resting cysts of G. catenatum in Tasmania suggest it was introduced to southern Tasmania during the early 1970's, possibly via ballast water from either Japanese, Korean or Spanish populations. The cysts of this species have now been widely reported, even from areas where G. catenatum has rarely, or never, been detected in the plankton. The reported cyst diameter varies considerably (17-63 p,m diameter), displaying a bimodal size distribution at some localities, suggesting that cysts may belong to a species complex of two or more related species. This work examines the distribution, and morphological and genetic variation, of the Gymnodinium catenatum species complex to resolve three distinct microreticulate cystforming species. Resolution of the species complex allowed the elucidation of population genetic relationships between strains of the toxic \true G. catenatum\" isolated from Japan Spain and Portugal and Australia to examine the hypothesis that Tasmanian G. catenatum was introduced to Australia (Tasmania) from one of these two potential source populations. Mapping the distribution of G. catenatum by examining sediments for the microreticulate cysts is hampered by their low abundance in coastal sediments and the low proportion of intact and viable specimens. Cyst concentration methods (sodium polytungstate density centrifugation) and PCR-based genetic identification methods were developed to improve cyst survey detection limits. Sediment surveys of 105 sampling sites at 17 estuarine and coastal locations demonstrated that microreticulate cysts are widely distributed in Australian coastal sediments. Two distinct morphotypes were noted: a \"small-form\" cyst (17-28 pm) widely distributed in temperate and tropical Australian estuaries and a \"large-form\" typical of G. catenatum ( 37-62 p.m) which was restricted to the coasts of south-eastern Tasmania southern Victoria Port Lincoln (South Australia) and the Hawkesbury Estuary (NSW). Germination experiments showed that the smaller cyst-type belonged to a species that was morphologically and genetically distinct from both G. catenatum and the European rnicroreticulate cyst species G. nolleri Ellegaard et Moestrup. This new species is described herein as Gymnodinium microreticulatum sp. nov. Bolch et Hallegraeff. The possible origin and evolution of G. catenatum was investigated by examining the genetic relationships among 27 species of gymnodinoid prorocentroid and peridinoid dinoflagellates using partial sequences of the large sub-unit ribosomal RNA gene. The phylogenies constructed confirmed earlier findings from small sub-unit (SSU) RNA studies. The relationships of the 21 free-living gymnodinoids correlated with their morphological and cytological features with the loop-apical-grooved species forming four clusters delineated by chloroplast structure and arrangement and resting cyst morphology. The G. catenatum complex (G. catenaturn G. nolleri and G. microreticulatum) formed a distinct monophyletic lineage arising from the base of this group. Within the complex G. microreticulatum diverged earliest followed by G. nolleri and G. catenatum. Comparison of nuclear volumes of the three species suggests that the group may have evolved by polyploidy from a G. microreticulatum-like ancestor. The abundant LSU-rDNA sequence variation among G. microreticulaturn isolates contrasted the absence of verifiable sequence variation among G. catenaturn from Australia Hong Kong Japan Spain and Uruguay Analysis of allozymes and large-subunit ribosomal RNA sequences failed to reveal genetic polymorphism among G. catenatum strains from Australia Japan Spain and Portugal. However reproductive compatibility analysis demonstrated extensive intrapopulation compatibility and an outbreeding multiple-group mating system. Mating success analysis (by cyst production) and variation in post-meiotic progeny viability from inter-population crossing experiments indicated that Japanese and Spanish strains were more closely related to each other than to Australian strains. Mating studies were supported by genetic studies using RAPD-PCR. Genetic variation was partitioned primarily within populations (87%) consistent with a sexually outbreeding species as confirmed by mating studies. The G. catenatum strains could be clearly separated into regional clusters: Australia Japan and Spain/Portugal. The Spanish/Portuguese and Japanese clusters were most closely related with the Australian cluster more distant and almost equally related to the others. The similarity between Japanese and Spanish G. catenatum compared to Australian strains suggests recent dispersal between these two populations. The source population for Australian G. catenatum remains unclear however the data support a secondary relocation of Tasmanian G. catenatum to mainland Australia possibly via a domestic shipping vector. Geographic and temporal clustering of Tasmanian strains by isolation location and bloom year indicates that genetic exchange between neighbouring estuaries is limited and that Tasmanian G. catenatum blooms are composed of localised estuary-bound sub-populations. Re-examination of the dinoflagellate fossil record in light of recent molecular phylogenetic data suggests that the G. catenatum complex evolved near the start of the Cretaceous period [circa 150 million years ago (Mya)]. Using a G. catenatum complex \"molecular clock\" based on LSU-rDNA sequences it is estimated that the common ancestor of the complex (probably a G. microreticulatum-like species) evolved around 130-140 Mya and that the G.nolleri and G. catenatum lineages diverged about 16-19 Mya. The known modern distributions of the three species suggest a European evolutionary origin of G. catenation followed by a geologically recent global dispersal. Considering the lack of rDNA variation among the five populations examined this dispersal is conservatively estimated to have occurred well within the last 25 thousand years. Whether dispersal has been by natural processes or assisted by human means (or a combination of both) is not yet clear. This study demonstrates that PCR-fingerprinting methods such as RAPD-PCR can discriminate fine-scale genetic population structure and discriminate global population clusters. Comparative genetic studies of more G. catenation populations would allow a better assessment of global relationships and may identify population genetic dines that correlate with G. catenatum natural dispersal corridors or discontinuities that imply trans-oceanic transfer and human introduction."