Classification and identification of Pfiesteria and Pfiesteria-like species.

Dinoflagellates can be classified both botanically and zoologically; however, they are typically put in the botanical division Pyrrhophyta. As a group they appear most related to the protistan ciliates and apicomplexans at the ultrastructure level. Within the Pyrrhophyta are both unarmored and armored forms of the dominant, motile flagellated stage. Unarmored dinoflagellates do not have thecal or wall plates arranged in specific series, whereas armored species have plates that vary in thickness but are specific in number and arrangement. In armored dinoflagellates, the plate pattern and tabulation is a diagnostic character at the family, subfamily, and even genus levels. In most cases, the molecular characterization of dinoflagellates confirms the taxonomy on the basis of external morphology; this has been demonstrated for several groups. Together, both genetic and morphological criteria are becoming increasingly important for the characterization, separation, and identification of dinoflagellates species. Pfiesteria and Pfiesteria-like species are thinly armored forms with motile dinospore stages characterized by their distinct plate formulae. Pfiesteria piscicida is the best-known member of the genus; however, there is at least one other species. Other genetically and morphologically related genera, now grouped under the common names of "Lucy," "Shepherd's crook," and cryptoperidiniopsoid, are being studied and described in separate works. All these other heterotrophic dinoflagellate groups, many of which are thought to be benign, co-occur in estuarine waters where Pfiesteria has been found.

Harmful algal bloom (HAB) species can produce a variety of biotoxins that can cause human illness, mass mortalities of aquatic organisms, and disease. Many of the biotoxins can be detected or assayed directly. In known toxin producers it is possible to determine when potentially harmful algal species are present in sufficiently high abundance levels to produce toxins, then to employ the more sensitive analytical techniques or bioassays to quantify the amount of toxin(s) present in the environment. Unfortunately, for some toxins there currently are no assays or appropriate analytical techniques because the toxins remain uncharacterized. In this situation the most efficient strategy for identifying a potential public health threat is to rapidly determine if high, potentially toxic concentrations of the organism are present so that appropriate response plans can be implemented. Quantifying the number of HAB species present in a sample can present difficulties because morphologically similar, but benign, species can be mistaken for the toxin-producing species. This is true for Pfiesteria piscicida, a small heterotrophic dinoflagellate that has been associated with fish kills and the production of a putative biotoxin in North Carolina and Maryland (1)(2)(3). Originally, it was assumed that small heterotrophic dinoflagellates of a particular shape, and approximately 10 µm in length, were all P. piscicida. That morphologically similar co-occurring species could be misidentified with P. piscicida has been recognized (4). Because a number of related nontoxic Pfiesteria-like species co-occur with P. piscicida, assessment of the potential public health threat of this species is complicated. This article is part of the National Conference on "Pfiesteria: From Biology to Public Health" and covers what is currently known about the higher-level placement of the Pfiesteria-like organisms and the taxonomic relationship between P. piscicida and co-occurring morphologically similar species.

Higher-Order Taxonomic Placement of P. piscicida and Related Dinoflagellates
Pfiesteria and several Pfiesteria-like genera are dinoflagellates classified in the botanical division Pyrrhophyta. They are morphologically and genetically related and may derive from a common ancestor. The only valid and available named species as of this conference (5) is P. piscicida Steidinger and Burkholder (6), which has a zoospore stage characteristic of armored dinoflagellates. A second species, to honor Dr. Sandra Shumway, was named after the conference (7). Armored dinoflagellates have an outer wall of segments, or plates, arranged in specific, mostly horizontal series (Kofoidian series) ( Figure 1). These series constitute a plate pattern and formula and are useful in separating genera (8). There is a good correlation between the phylogenetic relationships among dinoflagellates based on plate structure and those based on small subunit rRNA and rDNA sequences (9,10).
Because dinoflagellates have characteristics that historically have been considered "botanical," e.g., presence of chloroplasts, as well as "zoological," e.g., flagella and heterotrophic mode of feeding, there are several classification schemes. Recently, Fensome et al. (11) proposed a single classification scheme that addressed fossil and extant groups and followed the International Botanical Code of Nomenclature. In the same article, the authors reviewed 39 classification schemes for living dinoflagellates dating back to the 1800s. Some of the classification characteristics included plate tabulation, life cycle, and the unique dinokaryon nucleus. In that work (11) and elsewhere, plate tabulation is considered the most important morphological characteristic to differentiate within the division Pyrrhophyta at the family or subfamily rank. It is also used at the generic level in Dinoflagellates can be classified both botanically and zoologically; however, they are typically put in the botanical division Pyrrhophyta. As a group they appear most related to the protistan ciliates and apicomplexans at the ultrastructure level. Within the Pyrrhophyta are both unarmored and armored forms of the dominant, motile flagellated stage. Unarmored dinoflagellates do not have thecal or wall plates arranged in specific series, whereas armored species have plates that vary in thickness but are specific in number and arrangement. In armored dinoflagellates, the plate pattern and tabulation is a diagnostic character at the family, subfamily, and even genus levels. In most cases, the molecular characterization of dinoflagellates confirms the taxonomy on the basis of external morphology; this has been demonstrated for several groups. Together, both genetic and morphological criteria are becoming increasingly important for the characterization, separation, and identification of dinoflagellates species. Pfiesteria and Pfiesteria-like species are thinly armored forms with motile dinospore stages characterized by their distinct plate formulae. Pfiesteria piscicida is the best-known member of the genus; however, there is at least one other species. Other genetically and morphologically related genera, now grouped under the common names of "Lucy," "Shepherd's crook," and cryptoperidiniopsoid, are being studied and described in separate works. All these other heterotrophic dinoflagellate groups, many of which are thought to be benign, co-occur in estuarine waters where Pfiesteria has been found. most of the orders containing the majority of armored species, e.g., Gonyaulacales and Peridiniales (12).
In general, dinoflagellates are thought to have affinities with ciliates and apicomplexans at the ultrastructure level because of similar organelle structures such as tubular mitochondrial cristae, type of spindle, rod-shaped trichocysts, and cortical vesicles, as well as at the genetic level (9,11,13). Dinoflagellates distinctly have a dinokaryon nucleus with continually condensed chromosomes throughout the entire life cycle, or in at least one stage of the life cycle, and characteristic flagella insertion.
A possible phylogenetic scenario or tree was presented by Fensome et al. (11), with support from the fossil record. This proposed phylogeny suggests that, structurally, unarmored dinoflagellates with unfilled cortical vesicles (hundreds) preceded armored forms in the evolutionary process; the armored forms then evolved from groups having many plates (vesicles filled with polysaccharide plate material) to those having only a few plates, as in the order Prorocentrales. In this scheme, the placement of Pfiesteria and related genera depends on several attributes that need further clarification. Originally, Pfiesteria was placed in the order Dinamoebales, which incorporates species with multiple life history stages, including dominant amoeboid forms. However, there are concerns that the dominant life history stage is a motile dinospore or coccoid stage with closer morphological affinities to the order Peridiniales (14). Does Pfiesteria have dominant amoeba and/or coccoid stages? Even if the dominant state is an amoeba, there have been proposals to transfer Pfiesteria from the order Dinamoebales to the order Phytodiniales (11) or to place Pfiesteria specifically into the order Blastodiniales (9). Landsberg et al. (15) raised similar questions about life-cycle stages in the parasitic dinoflagellate Amyloodinium, and it was suggested that A. ocellatum belonged in the order Peridiniales on the basis of the morphology of the dinospore stage. Recent phylogenetic work on the rRNA ITS and 5.8S regions supports the placement of Pfiesteria in the Peridiniales or in a group between the Peridiniales and the Blastodiniales (16). All these options need to be carefully considered and, for now, most are theoretical schemes based on limited morphological, genetic, and physiological life history data for these orders.

Pfiesteria-Like Organisms
Single cells of flagellated motile stages of Pfiesteria and Pfiesteria-like organisms were isolated from water samples collected in    (17) were similarly isolated and maintained at other institutions. Clonal isolates were used to prepare specimens for scanning electron microscopy (SEM) so that the thecal plate patterns and tabulations could be determined reliably from one population. Over 63 clonal cultures were prepared for SEM, and they resulted in more than 1,000 micrographs of individual cells. In addition, more than 3,000 micrographs of individual cells were reviewed from field samples. The most frequently used fixation and preparation protocol was that of Truby (18), in which cells were either stripped or osmotically swollen to visualize sutures. Sutures ( Figure 1B) are the visible lines between plates, like grouting between tiles. These same isolates were used for genetic studies (9,(19)(20)(21)(22) and for the development of whole cell and other molecular probes.

Results and Discussion
All the Pfiesteria and Pfiesteria-like dinospores are heterotrophic, typically with a dinokaryon nucleus in the hypotheca and food vacuoles in the epitheca ( Figure 1A). The longitudinal flagellum has two parallel parts, and the peduncle, a feeding organelle, emerges from under the flexible right sulcal plate. The sulcus is not straight but offset. Food vacuoles can contain whole phytoplankton cells or whole chloroplasts. At least in P. piscicida, Lewitus et al. (23) documented the chloroplasts to be functional for a period of time. The genus Pfiesteria is characterized as having a plate formula/tabulation of apical pore complex (APC) (pore plate, closing plate, and X plate), 4´, 1a, 5´´, 6c, 4s, 5´´´, 0p, 2´´´´ (Figures 2A, 2B, 3). An unnamed marine Pfiesteria (Pfiesteria. sp. "marina," Figure 4) (24) has the same plate formula (Table 1). Both have a triangular 1a (first intercalary plate) ( Figures 1B,2B) on the left shoulder, which is one of the characteristics of the genus. Another proposed Pfiesteria (7), Pfiesteria shumwayae ( Figure 5), has a larger, almost rectangular 1a plate and a plate formula distinct from that of Pfiesteria (six precingular plates instead of five). This species, because of its difference in plate tabulation, should be reevaluated as to its genetic and morphologic affinity with Pfiesteria. Both species have been implicated as producing toxins (3,6,7), though a purified toxin has yet to be isolated and characterized for either species.
Pfiesteria-like organisms, by our definition, include numerous species that are superficially similar when viewed at the light microscopic level but distinct at the ultrastructural level. These are small (10-20 µm) cells lightly armored with defined plate formulae/tabulations and occupying ecological niches similar to Pfiesteria. Many of these species are not harmful but may be closely related genetically to those that are. The following summarizes the current knowledge regarding the taxonomic relationships of these similar groups and genera.
Another new heterotroph species in the order Peridiniales, with a common name, "Shepherd's crook," because of its unusual crook-shaped canal or X plate, co-occurs geographically with species in the family Pfiesteriaceae, but it does not have the appropriate plate formula or offset sulcus to be included in that family (Figure 8). The species needs further workup to properly assign it to a genus. Its potential to produce toxins is unknown, but the species is thought to be benign.
Pfiesteria and its relatives are widely distributed, and like other dinoflagellates, they can have benthic stages, which could account for their recurrence in specific areas. Burkholder and Glasgow (26) and colleagues (27) have found there are benign stages of P. piscicida and that when exposed to fish, the organism may become toxic and produce an ichthyotoxic compound. However, neither the morphological (4,6) nor the biochemical identification of species, using molecular probes (19,20,28), differentiates toxic from nontoxic or noninducible P. piscicida forms. Hence, from a public health standpoint, the presence of P. piscicida alone can potentially overestimate any toxic threat from this species. The same situation applies to P. shumwayae, as both species are part of the same toxic Pfiesteria complex (26).