Pfiesteria Piscicida and P

Journal of Experimental Marine Biology and Ecology 264Ž. 2001 29–45 www.elsevier.comrlocaterjembe

Chemosensory attraction of zoospores of the estuarine dinoflagellates, Pfiesteria piscicida and P. shumwayae, to finfish mucus and excreta

Paul J. Cancellieri, JoAnn M. Burkholder ), Nora J. Deamer-Melia, Howard B. Glasgow Center for Applied Aquatic Ecology, North Carolina State UniÕersity, 620 Hutton St., Suite 104, Raleigh, NC 27606, USA

Received 26 October 2000; received in revised form 13 November 2000; accepted 21 May 2001

Abstract

Toxic strains of the estuarine dinoflagellates, Pfiesteria piscicida and P. shumwayae, can cause fish death and disease, whereas other estuarine ‘lookalike’ species such as cryptoperidiniopsoids have not been ichthyotoxic under ecologically relevant conditions. Chemosensory attraction of three functional types of these Pfiesteria spp. were separately evaluated for their attraction to fresh fish mucus and excreta. Clonal cultures of actively toxicŽ. TOX-A, engaged in killing fish and temporarily nontoxicŽ tested as toxic but without access to live fish for )1 week to 5 months Ž.in most experiments, F3 months as ‘short-duration’ TOX-B; and without access to live fish for G1.5 years as ‘long-duration’ TOX-B. functional types of P. piscicida and P. shumwayae were derived from the same clones whereas the non-inducible culturesŽ NON-IND, tested as incapable of toxic activity in the presence of fish. , of necessity, were from different clonal isolates. NON-IND cultures previously had been grown on algal prey for 3–8 months, and had repeatedly been tested as incapable of causing fish distress, disease or death via toxic activity. Attraction to fish materials was based on the number of zoospores that entered microcapillary tubes containing sterile-filtered 15-ppt waterŽ. controls , vs. entry into tubes with sterile-filtered mucus and excreta Ž.collected in 15-ppt water that had been collected from live tilapia, bluegill, hybrid striped bass, and Atlantic menhadenŽ. tested separately within 3 h of removal from live fish . TOX-A zoospores of both Pfiesteria species exhibited the strongest attraction to the fish mucus and excreta, with comparable response to the materials from all four test fish species. Short-duration TOX-B zoospores showed an intermediate response that apparently depended on the duration of mucus separation from the live fish: the shorter the separation period, the stronger the zoospore attraction

) Corresponding author. Tel.: q1-919-515-2726; fax: q1-919-513-3194. E-mail address: joann– [email protected]Ž. J.M. Burkholder .

0022-0981r01r$ - see front matter q2001 Elsevier Science B.V. All rights reserved. PII: S0022-0981Ž. 01 00299-4 30 P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 to the fish materials. In contrast to TOX-A and short-duration TOX-B zoospores, NON-IND and long-duration TOX-B zoospores generally showed little or no response to the fish materials. Zoospores of the cryptoperidiniopsoid demonstrated a moderate attraction that did not appear to depend on the time of isolation from fish. TOX-A zoospores were also tested for attraction to sterile-filtered vs. non-filtered fish mucusŽ. time separated from the live animal, 3–96 h . These zoospores, which initially had been actively attracted, were no longer attracted to the unfiltered fish materials after 48 h, whereas attraction to the sterile-filtered fish mucus and excreta persisted throughout the duration of the experiment. Thus, the attractant signal in the materials was degraded or effectively blocked by the bacterial community within hours of isolation from live fish. This study indicates the importance of functional type or toxicity status, and the importance of the history of exposure to live fish, in the behavioral ecology of Pfiesteria spp. Initial attraction to fish materials strongly depended on the functional type, and on the history of toxic activity. Non-inducible and long-duration TOX-B cultures of Pfiesteria spp., unlike actively toxic and short-duration TOX-BŽ. potentially toxic strains, initially were virtually unresponsive to fish mucus. q 2001 Elsevier Science B.V. All rights reserved.

Keywords: Chemosensory; Functional type; Estuary; Fish; Mucus; Non-inducible; Toxic dinoflagellate; Toxic Pfiesteria complex; Cryptoperidiniopsoid

1. Introduction

Dinoflagellates within the genus Pfiesteria Žtwo species known thus far, P. piscicida Steidinger & Burkholder and P. shumwayae Glasgow and Burkholder. form the toxic Pfiesteria complexŽ. TPC; Glasgow et al., 2001; Burkholder et al., 2001 . Their ichthyotoxic strains have been implicated as primary and secondary causative agents of certain major fish kills and fish epizootic in the largest and second largest estuaries on the US mainlandŽ Chesapeake Bay and the Albemarle-Pamlico Estuarine System, respectively; Burkholder et al., 1992; Burkholder and Glasgow, 1997. . They are stimulated to produce toxin by live fish or their fresh secretarexcreta and tissues Ž.Burkholder and Glasgow, 1997 , to which they are strongly attracted Ž Pfiesteria Interagency Coordination Working Groupwx PICWG , 2001; Burkholder et al., 2001. . Toxic strains of Pfiesteria spp. have a complex life cycle including an array of flagellated, amoeboid, and cyst stagesŽ Burkholder and Glasgow, 1995; Steidinger et al., 1996. . The presence of live fish stimulates amoebae and cysts to produce toxic zoospores, and nontoxic zoospores to become toxicŽ Burkholder and Glasgow, 1995, 1997; Burkholder et al., 2001. . On the basis of repeated observations of Pfiesteria toxicity in the presence vs. absence of fishŽ. Burkholder et al., 1992, 1995; Burkholder and Glasgow, 1997 , three functional types of Pfiesteria zoospores have been definedŽ. Burkholder et al., 2001 . TOX-A zoospores are actively toxic, and are found in the presence of live fish Ž.Burkholder and Glasgow, 1997; PICWG, 2001 . Within hours of separation from fish prey, zoospore toxicity diminishes, then is re-induced after additional exposure to fish Ž.Burkholder and Glasgow, 1997 . TOX-B zoospores are capable of toxicity, but cease toxin production in the absence of live fish and prey on other organisms such as bacteria and algae. This functional type has been described as temporarily nontoxicŽ Burkholder and Glasgow, 1997.Ž , although it may retain negligible or residual toxicity Springer, P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 31

2000.Ž . As in many other toxic algal species here, including heterotrophic dinoflagellates; e.g., Anderson, 1990; Gentien and Arzul, 1990; Bates et al., 1998; Edvardsen and Paasche, 1998.Ž , non-inducible strains NON-IND functional type . of both Pfiesteria spp. have been isolated. These strains are apparently incapable of toxin production based on present knowledge about toxicity, although molecular controls that can re-initiate toxin production in this functional type may eventually be foundŽ Burkholder et al., 2001. . As a separate phenomenon and an artifact of culture conditions, toxic strains become non-inducible, or lose toxin-producing ability, after extended culture with fish Ž.typically, G6 months , accelerated when Pfiesteria is given other prey in the absence of fishŽ. toxicity loss in most clonal cultures within G6 weeks . Other researchers have described Pfiesteria lookalike speciesŽ term, APfiesteria-likeB as used by Steidinger et al., 1996; Litaker et al., 1999; Seaborn et al., 1999; Marshall et al., 2000. that have been informally named as ‘shepherd’s crook,’ ‘Lucie’, and Cryptoperidiniopsis gen. ined. Že.g., C. brodyii gen. et sp. ined.; Dr. K. Steidinger, Department of Environmental Protection, Florida Marine Research Institute, St. Petersburg, FL, USA, personal com- munication. . These organisms have been present in estuaries where Pfiesteria spp. have been documented. Unlike toxic strains of Pfiesteria spp., thus far none of the lookalike species have been ichthyotoxic under ecologically relevant conditionsŽ live or sonicated cells and fish in the standardized fish bioassay process, at similar densities as found under field conditions; Marshall et al., 2000; Burkholder et al., 2001. . Our research team has sought to compare the two known toxic Pfiesteria spp. and lookalike species by investigating their cellular and behavioral differences in response to various environmental stimuli. Toxic Pfiesteria outbreaks require that benthic and suspended stages of TPC species Ž.amoebae and cysts, and TOX-B zoospores, respectively producertransform to actively toxic zoosporesŽ. TOX-A functional type when sufficient live fish are detectedŽ Burk- holder and Glasgow, 1997; Burkholder et al., 1999. . TPC zoospores have shown positive chemotaxis toward fresh fish tissues in laboratory trials; therefore, it is expected that toxic strains of TPC zoospores exhibit positive chemotaxis toward fish and their fresh secreta in estuarine watersŽ. Burkholder and Glasgow, 1997 . The stimulatory substance in fish materials is believed to be a water-soluble chemical released by live fish, including potential candidates within methylamines, methylaminoxides, mucus proteins, or mucopolysaccharidesŽ Burkholder et al., 1995; Burkholder and Glasgow, 1997; Boriss et al., 1999. . Fresh fish secreta has been documented to affect aquatic microfauna other than dinoflagellates. For example, zooplankton species respond to glycosaminoglycans in fish mucus, which induce life cycle changes and migratory activity as defenses against the fish predatorsŽ. Forward and Rittschof, 1999, 2000 . In Pfiesteria, cessation of toxin production soon after fish deathŽ e.g., Burkholder and Glasgow, 1997. suggests that the fish signal is rapidly removed or ‘lost’ from the water. Aside from tests for toxicity by TOX-A zoospores in the presencerabsence of fish materials, little is known about comparative responses of the two known Pfiesteria species to fresh fish secretarexcreta, or behavioral differences in zoospore response depending on the fish species. The objectives of this research were to compare the initial behavior of TOX-A, TOX-B, and NON-IND P. piscicida and P. shumwayae vs. a cryptoperidiniopsoid species in response to a mixture of fish mucus and excreta; to test 32 P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 whetherrhow the behavioral responses differ depending on the fish species; to test whetherrhow the behavioral responses of TOX-B and NON-IND Pfiesteria differ depending on the duration away from live fish or re-exposure to live fish; and to assess the duration of the response by Pfiesteria zoospores when bacteria are present vs. absent in the fish-materials substrate. We hypothesized that TOX-A, TOX-B, and NON-IND zoospores would show progressively less attraction to fresh fish mucus and excreta, with TOX-A zoospores most strongly attracted and NON-IND zoospores least attracted. We also expected that TOX-B zoospores separated from live fish long-term would initially behave similarly toward fish as NON-IND zoospores. Considering information from other research indicating that NON-IND Pfiesteria becomes increasingly ‘plant-like’ or mixotrophic, with higher apparent reliance on kleptochloroplasts from algal preyŽ Par- row et al., in press. , we predicted that NON-IND isolates would show little or no attraction to fresh fish mucusrexcreta. Based on experience with cryptoperidiniopsoids, we anticipated that these isolates would show moderate to low attraction, comparable to the behavior of TOX-B or NON-IND Pfiesteria zoospores. We expected that comparison of the stability of the attractant signal in sterile-filtered vs. unfiltered fish materials over time would enable insights as to whether the substanceŽ. s in fish secretarexcreta that attracts TOX-A Pfiesteria is chemically degraded, or lost via microbial consump- tion.

2. Methods and materials 2.1. Pfiesteria and cryptoperidiniopsoid isolates A toxic culture of P. piscicida Ž.a271T-1 was isolated from the mesohaline Neuse Estuary, North CarolinaŽ. as in Glasgow et al., 2001 and cloned in uni-dinoflagellate culture with algal preyŽ cloned from CCMP757 Rhodomonas sp. culture from the Culture Collection for Marine Phytoplankton, Bigelow Laboratory, Bigelow, ME, USA. as a food source. Pfiesteria spp. have not yet been cultured successfully to yield high cell production without preyŽ Burkholder et al., 1992, 2001; Burkholder and Glasgow, 1995; Lewitus et al., 1995, 1999; Glasgow et al., 1998. , and it has not been possible to induce toxin production unless live fish are addedŽ. Burkholder and Glasgow, 1997 . Clonal Pfiesteria thus is defined as an isolate of uni-dinoflagellate P. piscicida or P. shumwayae, grown initially from one zoospore, with its associated endosymbiont bacteria that may be presentŽ. Steidinger et al., 1995; Lewitus et al., 1999 and its prey Ž.Burkholder et al., 2001 . The clone was isolated from water samples taken from fish-killing bioassaysŽ standardized fish bioassays of Burkholder et al., 1992, 1995, 2001; Burkholder and Glasgow, 1997; Marshall et al., 2000. using a flow cytometer ŽCoulterww EPICS Altrae with HyPerSorte system; Coulter Corporation, Miami, FL, USA.Ž equipped with a water-cooled INNOVAe Enterprise IIe Ion Laser Coherent, Santa Clara, CA, USA. , and excitation provided by a 150 mWr488 nm argon laser line. Sub-cultures of this isolate were used as one of four replicates each of TOX-A and TOX-B functional types of P. piscicida zoospores. The TOX-A P. piscicida was maintained with live fish preyŽ tilapia, Oreochromis mossambicus Peters, length 3–5 cm. for ca. 2 months after being isolated from natural phytoplankton samples, and prior P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 33 to the experiments. Toxicity of the clone was cross-confirmed by Dr. H. Marshall, Old Dominion University, Norfolk, VA, USAŽ. Marshall et al., 2000 . Assessment of toxicity by fish deathŽ using appropriately designed, standardized fish bioassays; Burkholder and Glasgow, 1997; Burkholder et al., 2001. in the presence of clonal Pfiesteria spp. has been used routinely in toxic Pfiesteria research because the toxinŽ. s of these dinoflagellates has been only partially characterizedŽ. Kimm-Brinson et al., 2001 . Thus, purified standards are not yet available to verify the reliability of assays that are being developed to detect Pfiesteria toxins in fish cultures and estuarine samplesŽ e.g., Fairey et al., 1999. . The clone was cross-confirmed as uni-dinoflagellate using a Heteroduplex mobility assayŽ Dr. D. Oldach, University of Maryland, Baltimore, MD; Oldach et al., 2000. . The identifications of Pfiesteria spp. from scanning electron microscopy of suture- swollen cellsŽ. Burkholder and Glasgow, 1995; Glasgow et al., 2001 were cross-checked usingŽ. 1 polymerase chain reactionŽ. PCR species-specific molecular probes of our laboratoryŽ.Ž. 18S rDNA ; 2 a fluorescent in situ hybridizationŽ. FISH probe of Dr. P. Rublee, UNC-GreensboroŽ Rublee et al., 1999; and cross-corroborated by Dr. P. Rublee, University of North Carolina-Greensboro, Greensboro, NC, USA.Ž. ; and 3 the heteroduplex mobility assay of Oldach et al.Ž. 2000 ; and cross-corroborated by Dr. D. Oldach, University of Maryland, Baltimore, MD, USA. . Unless otherwise noted, the TOX-B zoospores were cultured as sub-cultures of TOX-A P. piscicida after G1 month of fish-killing activity. The TOX-B zoospores were cleaned, re-cloned, and grown with cryptomonad prey. Three additional clonal cultures from the mesohaline Neuse Ž.aB93BN3, a101296, and a416T were used for replicate clones 2–4 of TOX-A and TOX-B P. piscicida, prepared as above. To gain insights about the importance of duration of separation from live fish in influencing the subsequent initial response toward fish, we grew the TOX-B cultures for different periods on algal prey prior to these experiments. We designated ‘short-duration’ TOX-B isolates as having been separated from live fish and fed algal prey for F5 months. Each of the P. piscicida TOX-B isolates was tested at several time intervals of isolation from live fish, including 0.2, 0.5, 2, and 5 months for a271T1; 1 and 1.25 months for a416T; 0.25 and 0.75 months for a101296; and 0.5, 1, 1.5, and 3 months for aB93BN3. All TOX-B isolates were re-tested with live fish in standardized fish bioassays at the end of these experiments, to ensure that they were still toxicŽ rather than non-inducible. strains. Four other Neuse clones Ž.a114160, a89B2, a140B, and a271A , isolated from the same location using algal assaysŽ. Burkholder and Glasgow, 1995; PICWG, 2001 , were necessary for NON-IND P. piscicida Žspecies identification cross-confirmed as above; separated from live fish for )28 months since isolation from the Neuse Estuary. . The NON-IND isolates were grown with Rhodomonas prey for ca. 3 months prior to the experiments. All four NON-IND clones were re-tested and confirmed as unable to kill fish in repeated, standardized fish bioassays 4 weeks prior to the experiments. For P shumwayae, the following cultures were available: four TOX-A culturesŽ from clones a104TPA, a101127, a121B2, and a101272. ; three ‘short-duration’ TOX-B cultures from the same clones, separated from live fish for 0.5–3.0 months; two ‘long-duration’ TOX-B cultures from different clones Ž.a270A2, a130T1 , separated 34 P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 from live fish long-term for 1.5 years; and one NON-IND culture Ž.a598A . The P. shumwayae cultures were isolated from the mesohaline Neuse Estuary using similar procedures as those given for P. piscicida isolates, and toxicity of P. shumwayae was cross-corroborated by Dr. H. Marshall. As for P. piscicida, the short-duration TOX-B P. shumwayae cultures were assayed at several periods of isolation from fish prey, including 0.5, 1.0, and 1.25 months for a104TPA; 0.75 and 3.0 months for a101127 and a121B2. All six TOX-B isolates were re-tested with live fish in standardized fish bioassays at the end of these experiments, to verify that they were still toxicŽ rather than non-inducible. strains. The NON-IND P. shumwayae culture had been grown on algal prey for 8 months, and was tested as incapable of causing fish distress, disease or death in repeated, standardized fish bioassaysŽ 10 weeks in duration, terminating 4 weeks prior to the experiments. . Eight clones of cryptoperidiniopsoids Ža508A, a543A, a547A, a548A, a599A, a1040C, a1039C, and a101240. were isolated from Neuse field samples and fish bioassays, and were cultured on cryptomonad preyŽ cloned from Rhodomonas CCMP a757. for 1 week to 15 months. All eight isolates were verified as cryptoperidiniopsoids using PCR primers specific to that groupŽ. Rublee et al., 1999 , although not specific to one species. Therefore, it is likely that more than one species of cryptoperidiniopsoid dinoflagellate was represented by these isolates.

2.2. Fish mucusrexcreta collection

Fish mucus and excreta was collected from live tilapia, Atlantic menhaden Ž BreÕoortia tyrannus Latrobe; from the mesohaline Neuse Estuary, North Carolina, USA. , hybrid striped bass Ž Morone saxatilis=M. chrysops Rafinesque; from the Pamlico Aquacul- ture Field Laboratory, NCSU, Aurora, NC, USA.Ž , and bluegill Lepomis macrochiris L.; from Mr. T. Shedd, US Department of Defense, Fort Detrick, MD. . All fish were juveniles with an average mass of 15 g, and allŽ. except for Atlantic menhaden had been fed Tetra Marinw fish food once daily prior to collecting excreta and mucus. The fish were rinsed in 15-ppt waterŽ made with deionized water and Instant Ocean salts, then filtered using 0.22 mm-porosity, Pall-Gelman Acrodiscw filters. . Fish were placed into a Ziplocw plastic bagŽ. one individual per bag with 50 ml of sterile-filtered 15 ppt water, and the bag was sealed. Mucusrexcreta production was stimulated by repeated gentle squeezing of the bagged fish; after 4 min, the fish were removed, and 30 ml of material from the bag were collected. Since the fish were completely submerged, the water contained both mucus and excreta. The medium was sterile-filtered immediately upon collectionŽ 0.22 mm-porosity, Pall-Gelman Acrodiscw filters, including filtration of the menhaden materials in the field, followed by transport to the laboratory on ice. or unfiltered depending on the experiment. The excretarsecreta samples were held at 4 8C in darkness until useŽ. within 3 h of collection from the live animals .

2.3. Assay for chemosensory attraction

Microcapillary tubes were prepared using a Flaming-Brown micropipette puller Ž.Model P-97, Sutter Instruments, Novato, CA to an aperture size of 30"2 mm P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 35

Žmean"1 standard errorwx SE , ns30; modified from Spero, 1985. . The tubes were filled by capillary action with the sterile-filtered fish mucusrexcreta solution, and one end was capped with inert Teflonw stopcock grease. We also tested Pfiesteria spp. general response to organic molecules, using microcapillary tubes filled with an amino acid standard solutionŽ mixture of 12 amino acids in 15-ppt seawater at 2.5 nM each, aNCI0180; Pierce Chemical, Rockford, IL, USA. or with a bovine serum albumin wxBSA solution Ž.aA3350; Sigma, St. Louis, MO, USA in 15-ppt water. These two solutions were tested for response of Pfiesteria spp. functional types and the algal-fed cryptoperidiniopsoid dinoflagellates at several concentrations each, spanning the range of amino acid levels measured in the Neuse EstuaryŽ total dissolved amino acids 25 mM to 2.5 nM; BSA 3–300 mgrml; Center for Applied Aquatic Ecology, NCSU, unpub- lished data. . Control microcapillary tubes were filled with sterile-filtered 15-ppt water. Each trial consisted of one microcapillary tube immersed into a small dish containing 4 ml of culture water that had been filtered with a 30-mm mesh to remove large contaminants. Activity at the tube aperture was observed using light microscopy ŽOlympus AX-70 research light microscope, 600=, water immersion objective; Olym- pus, Melville, NY, USA.Ž and videotaped for 10 min Fig. 1. . The videotaping equipment included a cooled-chip digital video cameraŽ. Optronics, Goleta, CA, USA , S-VHS video recorderŽ. model SVO-9500MD, Sony Electronics, Park Ridge, NJ, USA , and 53-cm video monitorŽ. Sony Electronics . Total cells entering and leaving the microcapillary tube over the course of the experimentŽ. to determine net cells entering were recorded. The average rate of entry per minute was also calculated over the duration of each trial. Results were adjusted to the appropriate scale by normalizing total cell density to the median value for all experimentsŽ 104 zoospores mly1. , and then adjusting the value of net cells entering accordingly. TrialsŽ. control and experimental were performed in triplicateŽ that is, three trials per isolate; also one to four isolates per

Fig. 1. Frames from videotapes of microcapillary tube assays, illustrating zoospore attraction to tubes filled with an ultra-filteredŽ. 0.22 mm-porosity solution of fish mucusrexcreta. Frame A represents non-attraction Ž.NON-IND Pfiesteria functional type; scale bars10 mm , and frame B shows moderate to strong attractant Ž.Žpositive chemosensory behavior TOX-A Pfiesteria zoospores; scale bars10 mm. . 36 P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 functional type for each speciesŽ. given above, reflecting availability , and eight to nine periods of TOX-B zoospore separation from live fish. . Assays on cryptoperidiniopsoid isolates were carried out using three trials per isolate, five periods of separation from live tilapia, and similar controls as for Pfiesteria spp. To test the influence of re-exposure to live fish on the response of TOX-B and NON-IND Pfiesteria, four isolates each of TOX-B P. piscicida Ža101301, a101302, a101295, a101296.Ž and of NON-IND P. piscicida a101298, a101299, a101300, a101593.Ž that had been grown on algal prey CCMP757. for 1–5 months and 10–30 months, respectively, were re-inoculated into fish bioassays. Note that these isolates were re-tested and confirmed as TOX-B or as NON-IND prior to the experiments. After 48 h of exposure, attraction to fish secretarexcreta was measured as aboveŽ three trials per isolate, control and experimental. . Introduction of fish prey to the cryptoperidiniopsoid isolates was repeatedly attempted Ž.ns3 trials per isolate for the eight isolates , with and without high abundance of available algal preyŽ 3000–30,000 Rhodomonas cells mly1 . , but poor growth resulted in insufficient cell densities for the microcapillary assays ŽF800 cells mly1 ; similar data reported by Marshall et al., 2000. .

2.4. Time series experiment

A time series experiment was completed to assess the duration of viability of the chemical signal attractant for TOX-A Pfiesteria spp. in fish materials. Tilapia

Fig. 2. Dilution series experiments to determine the relationship between zoospore density and net entry into microcapillary tubes filled with ultra-filteredŽ. 0.22 mm-porosity tilapia mucus and secreta, or filled with ultra-filtered 15-ppt mediumŽ. controls , including data for all of the dinoflagellate species tested. The strong linear trend indicates a non-biological relationship, justifying use of normalization to allow comparison of trials with isolates of varying densities. P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 37 mucusrexcreta was collected as above. Sub-samples were not filtered, for comparison with fish material that was sterile-filtered to remove the microbial community. We inspected the filtered fish material with light microscopy, and did not observe bacteria in either fresh or stained samplesŽŽ stained with acidic Lugol’s solution Vollenweider, 1974.. , held at 4 8C, and observed within 4 h at 400= , after examining 10–20 fields of view. In contrast, unfiltered material contained abundant bacteria Ž)200 cells per field of view, 400=.Ž. Attraction of TOX-A zoospores cultures a271T and a104TPA, P. piscicida and P. shumwayae, respectively.Ž. was assayed in triplicate using the microcapillary tube method at 12-h intervals over a 72-h period, based on preliminary trials which indicated a degradation time for unfiltered fish materials of 30–50 h. The duration over which the mucusrexcreta mixture attracted the zoospores was compared, to determine the potential influence of the microbial population in degradation of the chemical signal.

Fig. 3. Attraction of zoospores of Pfiesteria spp. to fish secretarexcreta from several species of estuarine fish. TOX-A zoosporesŽ taken from cultures that had been in actively toxic, fish-killing mode for G2 months. and short-duration TOX-B zoosporesŽ temporarily nontoxic, separated from live fish and algal prey for F3 months.Ž consistently showed significant initial attraction to fish mucusrexcreta Bsbluegill, v sAtlantic menhaden, lsstriped bass, 's tilapia.Ž relative to zoospore activity in controls `, with microcapillary tubes containing ultra-filtered 15-ppt water.Ž . In contrast, NON-IND non-inducible status, separated from live fish and fed algal prey for 3–8 months.Ž and long-duration TOX-B cultures separated from live fish and fed algal prey for 1.5 years. showed very limited initial attraction to the fish materials. Overall, the TOX-A isolates of both Pfiesteria spp. were significantly more strongly attracted to the fish materials than cultures previously fed algal preyŽ. TOX-B, NON-IND functional types . The data indicate a negative second-order trend for attraction of TOX-A and short-duration TOX-B zoospores to fish materials with increasing duration of separation from live fish; and the data also show no differential response of zoospores among the four species of fish testedŽ for P. piscicida, ns3 replicates per isolate and 4 isolates per functional type, with short-duration TOX-B only; for P. shumwayae, ns3 replicates per isolate, with 4 isolates each for TOX-A and short-duration TOX-B functional types, two isolates for long-duration TOX-B, and 1 NON-IND isolate; note that each isolate was tested for its response to the mucusrexcreta from each fish species. . 38 P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45

2.5. Statistical analyses

It was hypothesized that cell density affects net entry of zoospores into microcapillary tubes based on a non-biological, linear relationship. In preliminary experiments, net entry of zoospores was measured in a dilution series, in which all factors other than density were held constant. The data showed that for both Pfiesteria spp., with or without fish mucusrsecreta, density and net entry increased in a linear fashionŽ Fig. 2; ns3 trials dilutiony1 speciesy12 ; r s0.986–0.992. . To correct for density range due to isolate variability, all values for ‘net zoospores entering tube’ were normalized to the median cell densityŽ 104 zoospores mly1. , as follows: net cells entering tube =1=10 4 snormalized net zoospores entering tube. cell density= culture volume Multi-factor ANOVAŽ MANOVA; as0.05; dependent variablesnormalized tube entry values.Ž was performed on the dataset to identify variables of interest i.e., variables that accounted for variation in attraction; SAS Institute, 1997. . Net entry values from triplicate trials were averaged, and mean zoospore entry values were used for statistical comparisons. Comparisons were made between control and experimental values, and between treatments in experimental trials using a Student’s t-test. Differences were

Fig. 4. Attraction of zoospores of the two species of Pfiesteria and a cryptoperidiniopsoid to a mixture of fish mucus and excreta. All zoospores were exposed to mucusrexcreta of all four fish species, with exception of the cryptoperidiniopsoid, which was tested with only tilapia mucus. For all functional types, P. piscicida Ž.B and P. shumwayae Ž.v attraction to fish mucusrexcreta showed no difference between species. The cryptoperidiniopsoid Ž.l showed an intermediate response, relative to the seawater controls Ž.` , at all time intervals isolated from fish, here referring to initial sample collection since re-introduction of cryptoperidiniopsoid cultures to fish was unsuccessful. The history of recent exposure to fish did not significantly affect the attraction of the cryptoperidiniopsoid dinoflagellate to fish materials. P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 39 considered significant at p-0.05, and behavior was regarded as positive attraction only when a significant difference was found between controls and experimental trials.

3. Results

Statistical analysisŽ. using General Linear Model testing, SAS Institute, 1997 indicated that variations between isolates, dinoflagellate species, and fish species were not significant Ž.p)0.05 . TOX-A and short-duration TOX-BŽ separated from live fish and fed algal prey for F5 months prior to the experiments. Pfiesteria spp. zoospores exhibited a strong attraction to the microcapillary tubes filled with fish mucusrexcreta, relative to zoospore response to filtered-water controlsŽ. Figs. 1 and 3 . The attraction progressively diminished as isolates were separated from fish prey for increasing duration, with the data generally following a second-order decay function. The response of the algal-fed cryptoperidiniopsoids was intermediate, comparable to that of a ‘mid-

Fig. 5. Rate of entry per minute of TOX-A, TOX-B, and NON-IND functional types of Pfiesteria spp. zoospores into microcapillary tubes filled with fish mucusrexcreta over 10-min experimental trialsŽ means"1 SE; ns3 for each of 4 isolates of each functional type of P. piscicida wshort-duration TOX-B only, separated from live fish and grown on algal prey for F3 monthsx ; ns3 for each of 4 isolates of TOX-A and short-duration TOX-B P. shumwayae wthe latter, separated from live fish and grown on algal prey for F3 monthsxw ; ns3 for each of 2 isolates of long-duration TOX-B separated from live fish and grown on algal prey for 1.5 yearsx and 1 isolate of NON-IND P. shumwayae; note that in this graph, the latter three isolates are considered collectively within the ‘non-inducible’ category because their behavior was comparable. . The asymptotic shape of the curves for TOX-A and short-duration TOX-B Pfiesteria spp. zoospores indicates that the 10-min period was of sufficient length to observe the greatest differences among the functional types. The data indicate that short-duration TOX-B zoospores, separated from fish for 1 week to 3 months, were less strongly attracted to the fish materials initially than TOX-A zoospores, which had consistently been exposed to live fish. That is, isolation from fish preyŽ. and subsistence on algal prey apparently promotes gradual loss of chemosensory attraction to fresh fish mucus and excreta by toxic strains of Pfiesteria spp. 40 P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 duration’ TOX-B P. piscicida isolate that had been separated from live fish for 10 months. Time isolated from fish prey did not affect the attraction of the cryptoperidiniopsoids to fish excretarsecretaŽ. Fig. 4 . Least attraction was observed when NON-IND zoospores and ‘long-duration’ TOX-B zoosporesŽ the two P. shumwayae isolates that had been separated from live fish for 1.5 years. were exposed to microcapillary tubes filled with tilapia mucusrexcreta. No statistical differences were evident between the NON-IND and ‘long-duration’ TOX-B zoospores; yet, even at this low level, the attraction was significantly higher when compared with Pfiesteria behavior toward controls. There was no significant difference in attraction toward fish species. None of the concentrations of the amino acids or the BSA solution elicited a chemosensory response from any functional type of TPC zoospores or the cryptoperidiniopsoid, even at concentrations 10-fold higher than normal in situ levels in the Neuse. TOX-A Pfiesteria spp. zoospores entered the mucus-filled microcapillary tubes at a significantly higher rate than did other functional typesŽ. Fig. 5 . These actively toxic zoospores appeared to respond more rapidly to the chemical signal, and reached maximum rate of entry within 6 min. In contrast, NON-IND and long-duration TOX-B zoospores reached a maximal rate of entry into the microcapillary tubes within 8 min, and this rate of entry was only ca. one-fourth the rate of the TOX-A zoospores. The cryptoperidiniopsoid zoospores exhibited an intermediate in rate of entry that was approximately half that of TOX-A Pfiesteria spp. zoosporesŽ. Fig. 5 . Short-duration TOX-B and NON-IND P. piscicida isolates that had been re-intro- duced to fish prey 48 h before the experimental trials showed marginally increased

Fig. 6. The extended time series experiment, showing the change in net number of Pfiesteria spp. TOX-A zoospores attracted to sterile-filtered versus unfiltered fish mucus and excreta, in comparison to TOX-A zoospore response to control tubes filled with ultra-filtered 15 ppt water, over a 72-h periodŽ means"1 SE, ns3. . Attraction to unfiltered fish materials was lost by 48 h, whereas the filtered fish materials attracted zoospores throughout the 72-h duration. The data indicate that the microbial consortium consumed or degraded the attractant substanceŽ. s in the fish mucusrexcreta. P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 41 attraction to fish secretarexcreta. For the four TOX-B isolates, the average normalized net entry value increased from 65"5 cells miny1 Ž. mean"1 SE previously fed algal prey, to 81"7 cells miny1 when tested following re-exposure to live fish for 48 h. The attractant signal in tilapia mucusrexcreta continued to significantly stimulate TOX-A zoospores of P. piscicida and P. shumwayae for at least 72 h, if bacteria were removed from the fish materials immediately after collectionŽ. Fig. 6 . In contrast, unfiltered fish materials rapidly lost attractant capability within 48 h after removal from live fish, indicating that loss of the signal is related to the presence of the microbial community.

4. Discussion

In this study, P. piscicida and P. shumwayae actively toxicŽ. TOX-A zoospores showed strong attraction to the mucusrexcreta from all four test fish species, indicating that these dinoflagellates are positively chemotactic toward a substanceŽ. s commonly produced by fish. The lack of an observed responseŽ. by any functional type to the amino acid or BSA solutions suggests that the positive chemotaxis of TOX-A Pfiesteria toward fish materials is not simply a nutritional response to organic molecules but, rather, a more specific stimulation by a substance in fish mucusrexcreta. The attractant response to materials from Atlantic menhaden was comparable to, but not notably greater than, the attraction to mucusrexcreta from the other species tested, of interest since 90% of the fish killed in estuarine toxic Pfiesteria outbreaks have been menhadenŽ. Burkholder et al., 1995, 2001; Burkholder and Glasgow, 1997 . Other characteristics of menhaden—such as high production of oily materials, schooling behavior, and tendency to feed in poorly flushed shallows where there are abundant phytoplankton and, often, Pfiesteria ŽManooch, 1988; Noga, 1993; Burkholder et al., 1995; Burkholder and Glasgow, 1997. —probably enhance detection by these dinoflagellates. In addition, the unusually thin epidermal covering on the ventral surface of this speciesŽ Dr. J.M. Law, College of Veterinary Medicine, NCSU, personal communica- tion. could facilitate exposure to Pfiesteria toxins and toxin impacts on the underlying viscera. The observed decrease in attraction to fish mucusrexcreta by even the short-duration TOX-B zoospores indicates that cultured Pfiesteria spp. lose the ability to rapidly detect and respond to live fish after feeding for extended periods on non-fish prey. Subsequent re-exposure to fish prey appears to stimulate some degree of increased response to fish materials. Whether this phenomenon occurs in the natural environment is unknown, but if certain biochemical pathways interact with molecular controls to elicit the positive chemotactic response toward fresh fish substances, a lag period may be required for the pathways to be activated. These findings support earlier studies, which indicated that the history of Pfiesteria exposure to fish is an important factor affecting toxicity and other aspects of the nutritional ecology of toxic strainsŽ. Burkholder et al., 1998, 2001 . Moreover, strains in the natural environment that inherently lack toxin-producing capability, or potentially toxic strains that have been separated from fish for long periods, may not prey on juvenile fish when they initially become available, as 42 P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 suggested by the behavior of NON-IND Pfiesteria spp. and long-duration TOX-B zoospores in this study. Other research has demonstrated that NON-IND Pfiesteria responses to nutrient enrichment and algal prey are distinct from responses of recently toxic strains that had been separated from live fish for F5 monthsŽ Burkholder et al., 2001; Parrow et al., in press. . The consistent response of the cryptoperidiniopsoids to the fish excretarsecreta, and the low cell production of cryptoperidiniopsoids in the presence of live fish, indicates that their attraction is not linked to the recent availability of fish prey. This information is consistent with current understanding of these dinoflagellates and their role in fish kills. Our difficulty in culturing this organism with fish in the absence of abundant algal prey suggests that these cryptoperidiniopsoid isolates had limited capability to switch from algal to live fish prey. Cryptoperidiniopsoids have, however, been shown to attack and consume larval fishŽ. Burkholder et al., 2001 . Thus, the significant chemosensory response that we observed may reflect the presence of a general attractant in fish excretarsecreta for some dinoflagellates. The data from the time series experiment with sterile-filtered vs. unfiltered fish mucusrexcreta point to an important role of the co-occurring bacterial consortium in the rapid breakdown of the attractive signal. The labile nature of the signal apparently results from its degradation by primary consumers, suggesting that bacteria may factor significantly in the dynamics of toxic Pfiesteria outbreaks. As instructive information for laboratory studies, sterile filtration of fish substances can extend their utility in research to understand the mechanisms of Pfiesteria spp. stimulation by fish. This technique also could be of value in reducing variability among secretarexcreta sources by enabling use of one collected sample of fish substances in certain experiments lasting 1–3 days. Pfiesteria spp. are dinoflagellates with complex life cyclesŽ Burkholder and Glasgow, 1995; Steidinger et al., 1996; Burkholder et al., 2001; Glasgow et al., 2001. and complex behavior toward a wide array of prey and other environmental stimuli ŽBurkholder and Glasgow, 1997; Burkholder et al., 1998, 2001; Glasgow et al., 1998; Parrow et al., in press. . This study supports the current understanding of functional types of TPC speciesŽ Burkholder and Glasgow, 1997; PICWG, 2001; Burkholder et al., 2001; Parrow et al., in press. , and the importance of working with demonstrated, cross-corroborated toxic rather than NON-IND Pfiesteria spp. strains in research to understand the behavior, physiology, and ecology of thesewx toxic dinoflagellates. Although other research has demonstrated that many other so-called ‘toxic algae’ have strains with undetectable toxin productionŽ e.g., Anderson, 1990; Gentien and Arzul, 1991; Skulberg et al., 1993; Bates et al., 1998, 1999; Edvardsen and Paasche, 1998. , the potential for significant differences in the comparative ecology of toxic vs. non-inducible strains within a given ‘toxic algal species’ generally has been ignored. Such comparative information takes on greater significance, given that most commercially available cultures are not screened for toxicity but are used, nonetheless, to make broad inferences about the behavior and ecology of toxic species. We hypothesize that, as for Pfiesteria, use of verified toxic strains will eventually be recognized to be of critical importance in research to understand the ecology and physiology of many harmful algae. We also note that commercial algal culture collections are not properly equippedŽ with specially P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 43 designed biohazard III facilities. to maintain toxic Pfiesteria strains, and generally list Apfiesteria-likeB species. We caution against applying data from use of such cultures toward understanding the cell biology, behavior, and ecology of TPC species, on the basis of this and other studiesŽ. Burkholder et al., 2001 .

Acknowledgements

Funding support for this research was provided by the US Environmental Protection Agency, an anonymous foundation, the North Carolina General Assembly, and the Department of Botany at NCSU. We thank A. Hamilton and N. Leaver of the Center for Applied Aquatic Ecology, NCSU for technical support, and M. Parrow and J. Springer for their counsel. N. Allen, D. Kamykowski, M. Mallin, S. Shumway, and two anonymous colleagues kindly reviewed the manuscript. []SS

References

Anderson, D.M., 1990. Toxin variability in Alexandrium species. In: Graneli,´ E., Sundstroem, B., Edler, L., Anderson, D.M.Ž. Eds. , Toxic Marine Phytoplankton. Proceedings of the Fourth International Conference on Toxic Marine Phytoplankton, Elsevier, New York, pp. 41–51. Bates, S.S., Garrison, D.L., Horner, R.A., 1998. Bloom dynamics and physiology of domoic-acid-producing Pseudo-nitzschia species. In: Anderson, D.M., Cembella, A., Hallegraeff, G.M.Ž. Eds. , Physiological Ecology of Harmful Algae. NATO ASI Series G: Ecological Sciences, vol. 41, Springer, Berlin, pp. 267–292. Bates, S.S., Hiltz, M.F., Leger, C., 1999. Domoic acid toxicity of large new cells of Pseudo-nitzschia multiseries resulting from sexual reproduction. Proceedings of the Sixth Canadian Workshop on Harmful Marine Algae, vol. 2261, pp. 289–301. Boriss, H., Boersma, M., Wiltshire, K.H., 1999. Trimethylamine induces migration of waterfleas. Nature 398, 382. Burkholder, J.M., Glasgow, H.B., 1995. Interactions of a toxic estuarine dinoflagellate with microbial predators and prey. Arch. Protistenkd. 145, 177–188. Burkholder, J.M., Glasgow, H.B., 1997. Pfiesteria piscicida and other Pfiesteria-like dinoflagellates: behavior, impacts, and environmental controls. Limnol. Oceanogr. 42, 1052–1075. Burkholder, J.M., Noga, E.J., Hobbs, C.W., Glasgow, H.B., Smith, S.A., 1992. New ‘phantom’ dinoflagellate is the causative agent of major estuarine fish kills. Nature 358, 407–410. Burkholder, J.M., Glasgow, H.B., Hobbs, C.W., 1995. Fish kills linked to a toxic ambush-predator dinoflagellate: distribution and environmental conditions. Mar. Ecol. Prog. Ser. 124, 43–61. Burkholder, J.M., Glasgow, H.B., Lewitus, A.J., 1998. Physiological ecology of Pfiesteria piscicida with general comments on ‘ambush-predator’ dinoflagellates. In: Anderson, D.M., Cembella, A., Hallegraeff, G.M.Ž. Eds. , Physiological Ecology of Harmful Algae. NATO ASI Series G: Ecological Sciences, vol. 41, Springer, Berlin, pp. 175–191. Burkholder, J.M., Mallin, M.A., Glasgow, H.B., 1999. Fish kills, bottom-water hypoxia, and the toxic Pfiesteria complex in the Neuse River and Estuary. Mar. Ecol. Prog. Ser. 179, 301–310. Burkholder, J.M., Glasgow, H.B., Deamer-Melia, N.J., 2001. Overview and present status of the toxic Pfiesteria complex. Phycologia 40, 106–134. Edvardsen, B., Paasche, E., 1998. Bloom dynamics and physiology of Prymnesium and Chrysochromulina. In: Anderson, D.M., Cembella, A., Hallegraeff, G.M.Ž. Eds. , Physiological Ecology of Harmful Algae. NATO ASI Series G: Ecological Sciences, vol. 41, Springer, Berlin, pp. 193–208. 44 P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45

Fairey, E.R., Edmunds, J.S., Deamer-Melia, N.J., Glasgow, H.B., Johnson, F.M., Moeller, P.D.R., Burkholder, J.M., Ramsdell, J.S., 1999. Reporter gene assay for fish killing activity produced by Pfiesteria piscicida. Environmental Health Perspectives 107, 711–714. Forward, R.B., Rittschof, D., 1999. Brine shrimp larval photoresponses involved in diel vertical migration: activation by fish mucus and modified amino sugars. Limnol. Oceanogr. 44, 1904–1916. Forward, R.B., Rittschof, D., 2000. Alteration of photoresponses involved in diel vertical migration of crab larva by fish mucus and degradation produces of mucopolysaccharides. J. Exp. Mar. Biol. Ecol. 245, 277–292. Gentien, P., Arzul, G., 1990. Exotoxin production by Gyrodinium cf. aureolum Ž.Dinophyceae . J. Mar. Biol. Assoc. U. K. 70, 571–581. Glasgow, H.B., Lewitus, A.J., Burkholder, J.M., 1998. Feeding behavior of the ichthyotoxic estuarine dinoflagellate, Pfiesteria piscicida, on amino acids, algal prey, and fish vs. mammalian erythrocytes. In: Reguera, B., Blanco, J., Fernandez, M.L., Wyatt, T.Ž. Eds. , Harmful Microalgae. Proceedings, VIIIth International Conference on Harmful Algal Blooms, Xunta de Galicia and Intergovernmental Oceano- graphic Commission of UNESCO, Paris, pp. 394–397. Glasgow, H.B., Burkholder, J.M., Springer, J.J., Morton, S.L., 2001. A new species of ichthyotoxic Pfiesteria. Phycologia 40, 60–71. Kimm-Brinson, K.L., Moeller, P.D.R., Barbier, M., Glasgow, H.B., Burkholder, J.M., Ramsdell, J.S., 2001.

Identification of a P2X7 receptor in GH4C1 rat pituitary cells: A target for a bioactive substance produced by Pfiesteria piscicida. Environmental Health Perspectives 109, 457–462. Litaker, R.W., Fester, P.A., Colorni, A., Levy, M.G., Noga, E.J., 1999. The phylogenetic relationship of Pfiesteria piscicida, cryptoperidiniopsoid sp., Amyloodinium ocellatum, and a pfiesteria-like dinoflagellate to other dinoflagellates and apicomplexans. J. Phycol. 35, 1379–1389. Lewitus, A.J., Jesien, R.V., Kana, T.M., Burkholder, J.M., Glasgow, H.B., May, E., 1995. Discovery of the AphantomB dinoflagellate in Chesapeake Bay. Estuaries 18, 373–378. Lewitus, A.J., Glasgow, H.B., Burkholder, J.M., 1999. Kleptochloroplastidy in the toxic dinoflagellate Pfiesteria piscicida Ž.Dinophyceae . J. Phycol. 35, 303–312. Manooch III, C.S., 1988. Fisherman’s Guide–Fishes of the Southeastern United States. North Carolina Museum of Natural History, Raleigh, North Carolina, 362 pp. Marshall, H.G., Gordon, A.S., Seaborn, D.W., Dyer, B., Dunstan, W.M., Seaborn, M., 2000. Comparative culture and toxicity studies between the toxic dinoflagellate, Pfiesteria piscicida and a morphologically similar cryptoperidiniopsoid dinoflagellate. J. Exp. Mar. Biol. Ecol. 255, 5–74. Noga, E.J., 1993. Fungal diseases in marine and estuarine fishes. In: Couch, A., Fournie, J.W.Ž. Eds. , Pathobiology of Marine and Estuarine Organisms. CRC Press, Boca Raton, pp. 85–100. Oldach, D., Delwiche, C., Jakobsen, K., Tengs, T., Brown, E., Kempton, J., Schaefer, E., Bowers, H., Glasgow, H.B., Burkholder, J.M., Steidinger, K.A., Rublee, P.A., 2000. Heteroduplex mobility assay guided sequence discovery: elucidation of the small subunitŽ. 18S rDNA sequence of Pfiesteria piscicida from complex algal culture and environmental sample DNA pools. Proc. Nat. Acad. Sci.Ž. U.S.A. 97, 4304–4308. Parrow, M.W., Glasgow, H.B., Burkholder, J.M., Zhang, C., 2001. Comparative response to algal prey by Pfiesteria piscicida, Pfiesteria shumwayae sp. nov., and an estuarine ‘lookalike’ species. In: Hallegraeff, G.M., Blackburn, S., Bolch, C., Lewis, R.Ž. Eds. Harmful Algal Blooms. IOC UNESCO, ParisŽ. in press . Pfiesteria Interagency Coordination Working Groupwx PICWG , 2001. Glossary of Pfiesteria-related Terms. U.S. Environmental Protection Agency, Baltimore,Ž available at http:rrwww.redtide.whoi.edurpfiesteriar documentsrglossary.html. . Rublee, P.A., Kempton, J., Schaefer, E., Burkholder, J.M., Glasgow, H.B., Oldach, D., 1999. PCR and FISH detection extends the range of Pfiesteria piscicida in estuarine waters. Va. J. Sci. 50, 325–336. Seaborn, D.W., Seaborn, A.M., Dunstan, W.M., Marshall, H.G., 1999. Growth and feeding studies on the algal feeding stage of a Pfiesteria-like dinoflagellate. Va. J. Sci. 50, 337–345. Skulberg, O.M., Carmichael, W.W., Codd, G.A., Skulberg, R., 1993. Taxonomy of toxic Cyanophyceae Ž.cyanobacteria . In: Falconer, I.R. Ž. Ed. , Algal Toxins in Seafood and Drinking Water. Academic Press, New York, pp. 1–28. Spero, H.J., 1985. Chemosensory capabilities in the phagotrophic dinoflagellate, Gymnodinium fungiforme.J. Phycol. 21, 181–184. P.J. Cancellieri et al.rJ. Exp. Mar. Biol. Ecol. 264() 2001 29–45 45

Springer, J.J., 2000. Interactions between two commercially important species of bivalve molluscs and the toxic estuarine dinoflagellate, Pfiesteria piscicida. Master of Science thesis, Department of Marine, Earth and Atmospheric Sciences, North Carolina State University, Raleigh, NC. 80 pp. Steidinger, K.A., Truby, E.W., Garrett, J.K., Burkholder, J.M., 1995. The morphology and cytology of a newly discovered toxic dinoflagellate. In: Lassus, P., Arzul, G., Erard, E., Gentien, P., Marcaillou, C. Ž.Eds. , Harmful Marine Algal Blooms, Technique et Documentation. Lavosier, Intercept, Paris, pp. 83–88. Steidinger, K.A., Burkholder, J.M., Glasgow, H.B., Hobbs, C.W., Garrett, J., Truby, E., Noga, E.J., Smith, S.A., 1996. Pfiesteria piscicida gen. et sp. nov.Ž. Pfiesteriaceae Fam. nov. , a new toxic dinoflagellate with a complex life cycle and behavior. J. Phycol. 32, 157–164. Vollenweider, R.A.Ž. Ed. , 1974. A Manual on Methods for Measuring Primary Production in Aquatic Environments. 2nd edn. Int. Biol. Program Handbook, vol. 12, Blackwell, Oxford, 225 pp.