Cultures. The Cryptomonas algal prey (HP 9101) was obtained from D. Stoecker (University of Maryland, Cambridge) and fed to Pfiesteria at 2- to 3-day intervals (≈10 algal prey per dinoflagellate) for 6 weeks. Algal prey and algae-fed Pfiesteria were grown in media made with natural seawater (Atlantic Ocean, 2 km offshore from Wilmington, NC; 0.22 m m-filtered, 15 practical salinity units). Fish prey and fish-fed Pfiesteria were maintained in media made with Coral Life or Instant Ocean salts [filtered (0.22-m m porosity), adjusted to 15 practical salinity units with deionized water].

Water Quality. General environmental conditions in the SFBs are summarized in Table 2. Presumptive Vibrio spp. densities were quantified as colony-forming units (cfu) from 0.05-0.10 ml inoculated onto thiosulfate-citrate-bile salts-sucrose (TCBS) agar (1, 2), normalized to cfu/ml in calculations. Plates were incubated at 35°C for 18-36 h, and they were held for 48 h before evaluated as negative.

Feeding Controls. Cultures used as feeding controls in SFBs are summarized in Table 3. These clones had been maintained on cryptomonad algal prey for months to years before SFBs.

Fish Histopathology. Tilapia (Oreochromis nilauticus, 12 juveniles, total length 6-7 cm; and 20 juveniles, total length 3-6 cm) were examined after exposure for 1-12 h to actively toxic P. shumwayae culture and culture filtrate (0.22-m m porosity), and they were compared with control fish that had not been exposed to toxic P. shumwayae. The P. shumwayae isolate (CAAE101272 = CAAE1024C, isolated from the Neuse Estuary in summer 2000, >1.00 × 10³ toxic zoospores per ml) had been lethal to fish in £ 4 h in culture before the pathology experiment. Fish were examined at 3-h intervals for evidence of toxic or other injury to the skin and other tissues by histopathology. At each interval, replicate subsamples of fish were carefully removed from the culture vessel, killed by an overdose of MS-222 [tricaine methanesulfonate, Argent Laboratories (3)], and examined for any grossly visible lesions. Multiple cross sections of each fish were prepared with a sharp necropsy blade, taking special care to avoid artifacts to the epidermis. Tissue samples were fixed in 10% neutral buffered formalin for 48 h, demineralized in 10% formic acid solution for 24 h, routinely processed, embedded in paraffin, sectioned at 5 m m, stained with hematoxylin and eosin (HE) (3), and examined by light microscopy by a single pathologist (J.M.L.). Other tissues were also examined, including the gills, and no remarkable microscopic abnormalities were found.

PCR/DGGE Microbial Analyses. Bacterial 16S rDNA was amplified by PCR in 50-m l reaction mixtures containing 10-20 ng of template DNA, 0.2 mM dNTPs, 2.5 mM MgCl₂, 1´ polymerase buffer (Sigma), 0.25 unit of Jump Start Taq DNA polymerase (Sigma), and 0.2 m M each of eubacterial primers 338F-GC (5'-CGCCCGCCGCGCCCCGCGCCCGTCCCGCCGCCCCCGCCCTCCTACGGGAGGCAGCAG-3') and 519RC (5'-ATTACCGCGGCTGCTGG-3') (4). The PCR mixtures were incubated at 94°C for 2 min, followed by 20 touchdown cycles of 30 sec at 94°C, 30 sec at 65°C, decreasing by 0.5°C each cycle, and 1 min at 72°C. The reaction continued for an additional 8 cycles of 30 sec at 94°C, 30 sec at 56°C, and 1 min at 72°C, followed by a 5-min extension at 72°C. For the bacterial community analysis, 10-20 m l of the PCR products were loaded directly onto an 8% polyacrylamide gel with a gradient of 30–65% denaturant (100% is 7 M urea/40% formamide). Electrophoresis was carried out for 5.5 h at 130 V at 60°C, using a Dcode Universal Mutation Detection System (Bio-Rad). DGGE gels were stained with ethidium bromide (0.5 mg/ml) and visualized on a UV transilluminator. The image was captured by using the AlphaImager System (Alpha Innotech, San Leandro, CA) and edited for size by using Photo Editor (Microsoft).

Evaluation of Bacteria- and Fungus-free Pfiesteria Cultures. These tests were conducted at 23°C and a 12:12-h light:dark cycle was imposed with cool white fluorescent lamps (50 m mol of photons per m² per sec). Agar for the plate assays was prepared as follows: 27.6 g of Difco 2216 marine agar per liter (Becton Dickinson & Company), agar (7.5 g of agar per liter), Bactopeptone (5 g/liter), and yeast extract (1 g/liter in deionized water).

PfTx: Purification, Chromatograms, and Cytotoxicity Screening. Dinoflagellate and algal cultures were harvested by filtering (Gelman no. 11873, thick glass fiber, Pall Corporation; or Whatman GF/C glass fiber filters; filters were frozen at –80°C until analysis within £ 2 months) to remove cell mass and physical debris from the culture media. Sample solutions were passed through a large glass column charged with bulk C₁₈ (10 m m, Phenomenex Aqua C18) to trap the organics from the media and to desalt the extract. Elution was carried out with 4 liters of methanol (MeOH) + 4 liters of ethyl acetate (EtOAc) to increase recovery (5). The material was loaded onto a glass column charged with Iatrobead silica (Iatron Laboratories), and active, semipurified PfTx was eluted by using MeOH after successive elutions with less polar solvents. The solution was rotary-evaporated to dryness at 45°C. Three successive desalted active fractions were produced, as defined by NMR spectroscopy (NMR) and mass spectrometry: (i) The residue placed on a size-exclusion column (HW40F, Tosohaus) and eluted with 4:1 water/acetonitrile. The NMR spectrum of the active fraction showed typical ¹³C resonances for glycosides present in the mixture at this step of the purification process. (ii) The active fraction was then passed through a reverse-phase, high-performance liquid chromatography (HPLC) scheme using three sequential C₁₈ columns (Phenomenex Aqua C18, Luna C18, Aqua C18). This step was designed to ensure that all free metals and remaining residual salts eluted in the water hold-time initially in the HPLC run. It yielded a much cleaner NMR spectrum, but loss of activity and a reduced purified residue. (iii) The active fraction was then passed through a bidentate C₁₈ column (Cogent bidentate reverse phase, MicroSOLV Technology, Long Branch, NJ), using the unique polar binding capacity of this column for further purification (Fig. 4b). Unless otherwise indicated (Fig. 4), the system was run at a flow rate of 0.5 ml/min with sample injection of 100 m l. PfTx was monitored at 254 and 272 nm. UV absorbance was measured as an analog electrical signal (in millivolts) and converted to milli-absorbance units (mAUs) by data acquisition software. It should be noted that fish toxicity has sometimes been reported from media made with synthetic seawater containing lipophilic phthalate esters (5). Synthetic media were used with fish-fed, but not algae-fed, clones in this study. The active fraction containing the hydrophilic PfTx was found in all toxic Pfiesteria clones but was not found in non-Pfiesteria controls containing synthetic or natural seawater-based media.

The toxic fractions were verified with the GH₄C₁ rat pituitary colorimetric assay (6). As previously described (5, 6), in vitro toxicity screening of various fish and mammalian cell lines has indicated that GH4 mammalian (rat) pituitary cells are highly sensitive to the hydrophilic bioactive substance from toxic Pfiesteria. We used this well defined in vitro cytotoxicity system, following ref. 6: Cytotoxicity was measured from microtiter plates by using the mitochondrial indicator 3-(4,5-dimethylthiazol-2,5-diphenyl) tetrazolium bromide (MTT) for endpoint measurement. GH₄C₁ rat pituitary cells (30,000 per well) were plated in 0.1 ml of appropriate medium in 96-well tissue culture plates (Corning Costar). Test samples (2 m l of methanolic extract) were added to each well and incubated for ≈24 h. Each fraction was analyzed in duplicate with a 2-m l MeOH negative control used to test the sample vehicle. After incubation, 15 m l of MTT (5 mg/ml in PBS) was added to each well and incubated for 4 h at 37°C. Mitochondrial dehydrogenases in live cells convert the MTT to an insoluble formazan crystal, resulting in a purplish color. The crystals were solubilized by adding 1% SDS/0.1 M HCl, and absorbance was read at 570 nm with a Titer Tek 96-well plate reader (EFLAB, Helsinki, Finland). The plate reader subtracts nonspecific absorbance by medium and nonconverted MTT to yield a corrected absorbance value.

Chromatogram plots and peak elution time were used to determine toxin elution time. Using bidentate column purification methods and matrix-assisted laser desorption/ionization (MALDI) time-of-flight (TOF) mass spectrometry (7, 8), we determined a molecular mass of ≈311 (MH⁺) Da for the PfTx active fraction. ¹³C NMR spectra repeatedly defined an aliphatic spin system with coupling to one heteroatom. Mass analysis of this fraction did not provide a definable molecular ion; the NMR data defined what would have been expected to be a lipophilic compound, yet it remained water soluble. Therefore, the active fraction was analyzed by inductively coupled plasma (ICP) mass spectroscopy (MS) to determine whether a metal was bound to the organic portion that had been observed with ¹³C NMR. ICP-MS results showed significant amounts of iron and copper bound to the organic ligand, which explains the more polar nature of this toxin that was not evident by cursory evaluation of the ¹³C NMR spectra. The ICP-MS data also showed substantial sulfur remaining in the purified fraction, which may define the heteroatom linkage to the metal as observed in the ¹³C NMR spectra The active fraction was earlier reported to be highly unstable, as has been found for other aliphatic ligated metal species in both oxidative and photodegradative processes (9-11).

Thus, the metallated complexes that form PfTx are highly unstable and exist in multiple congeneric/metallated forms as shown by ICP-MS data. These complexes are highly photosensitive and easily oxidized, explaining the rapid loss of activity and structural signals commonly observed with PfTx in NMR. The presence of these metals also explains the difficulty in obtaining meaningful mass spectral data, which has hampered analytical detection/characterization. Active fractions from P. piscicida and P. shumwayae toxic strains yield very similar ¹³C NMR spectra, supporting our earlier findings (12) that the two species produce PfTx.

FMAs, Assays with Extracts, and Comparison with Other Algal Toxins. FMAs with Pfiesteria cells were conducted at 20°C, 12:12-h light:dark cycle, ≈50 m mol of photons per m² per sec). Assays with each purified toxin extract, or with extracts from control cultures, were conducted in 24-well culture plates (Costar, Corning) containing 2 ml of salinity 15 filtered seawater (0.45-m m porosity; 1 juvenile C. variegatus per well). The data were comparable to published data for other toxic algae (Table 4).

Pfiesteria Abundance During Estuarine Fish Kills. A mixed clone culture of Pfiesteria piscicida (CAAE2200, CCMP1832) was used in developing the standard curve for semiquantitative PCR. The mixed clonal culture was fixed in 3% acidic Lugol’s solution, and 4- to 5-fold serially diluted in 6-ml aliquots. Subsamples (1 ml) were removed from each dilution set and cells were quantified by using a 0.5-mm hemocytometer on a Nikon Labophot-2 compound microscope. Four to five internal spikes were also aliquoted in equivalent volumes (6 ml) and cell densities were quantified. Two phytoplankton samples collected during estuarine fish kills [28 July 1998, Carolina Pines, Neuse Estuary, NC; 19 July 2000, Arnell Creek, DE (13, 14)] were dispensed in 6-ml volumes in triplicate, similarly as culture standards and internal spikes. DNA from the standards and internal spikes was extracted in 5-ml volumes by using a Purgene extraction system (Gentra Systems, Minneapolis) and following standard protocols.

Real-time PCR assays for P. piscicida and P. shumwayae (15) were then conducted by using the SmartCycler System (Cepheid, Sunnyvale, CA). Reaction protocols for each assay consisted of three stages, including an initial denaturation at 96°C for 75 sec followed by 50 cycles at 96°C for 5 sec and 62°C for 40 sec. Reaction mixtures for each assay consisted of two flanking primers and internal fluorescent-labeled probe. Internal probes were labeled on the 5' ends with 5-carboxyfluorescein and on the 3' ends labeled with 5-carboxytetramethylrhodamine (QIAGEN Operon, Alameda, CA). The following reagents were added: primers and internal probe at a final concentration of 0.2 m M; PCR buffer at a final concentration of 1´ (Promega); MgCl₂ at a final concentration of 2.5 mM; a dNTP mixture at final concentrations of 0.2 mM each (Stratagene); BSA at a final concentration of 10 m M; and Taq DNA polymerase at a final concentration of 0.2 unit/m l (Promega). Final reaction volumes were 25 m l. This semiquantitative PCR technique provided an approximate measure of P. piscicida abundance, considering that the number of 18S gene copies per cell was unknown, and that 18S gene copy number can vary with growth phase.

A linear relationship was obtained between standard and internal spike samples (Fig. 7a). The deviation from PCR run-to-run was minimal for the internal spike samples and unknowns, except for the 3,245 cells per ml internal spike, which fell slightly outside the standard curve (Fig. 7b). Estimates for the fish kill samples varied by ≈10% (DE sample, 490 and 548 P. piscicida cells per ml; NC sample, 288 and 320 P. piscicida cells per ml). These data provided by using semiquantitative PCR likely underestimated actual Pfiesteria cell densities in the acidic Lugol’s-preserved estuarine water samples that had been held 4-6 years (15). Values from semiquantitative PCR may also have been underestimated due to PCR inhibition from humic matter or other interfering substances in environmental samples. The Pfiesteria-like cell numbers reported during the two fish kills were higher, ≈3.00 × 10⁴ cells per ml [DE kill (14) and £ 1.4 × 10³ cells per ml (NC kill (13)]. The estimated cell densities are within the range of toxic P. piscicida cell densities that have been linked to fish death in estuarine fish kills [as Pfiesteria-like cells (13)] and SFBs (ref. 16 and this study).

1. Baumann, P. & Schubert, R. H. W. (1984) in Bergey’s Manual of Systematic Bacteriology, eds. Hold, J. G. & Krieg, N. R. (Williams & Wilkins, Baltimore), Vol. 1, pp. 516-550.

2. West, P. A. & Colwell, R. R. (1984) in Vibrios in the Environment, ed. Colwell, R. R. (Wiley, New York), pp. 285-363.

3. Leatherland, J. F. & Woo, P. T. K., eds. (1998) Non-infectious Disorders, Fish Diseases and Disorders (CABI, New York), Vol. 2.

4. Muyzer, G., DeWall, E. C. & Uitterlinden, A. G. (1993) Appl. Environ. Microbiol. 59, 695-700.

5. Moeller, P. D. R., Morton, S. L., Mitchell, B. A., Sivertsen, S. K., Fairey, E. R., Mikulski, T. M., Glasgow, H., Deamer-Melia, N. J., Burkholder, J. M. & Ramsdell, J. S. (2001) Environ. Health Perspect. 109, 739-743.

6. 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) Environ. Health Perspect. 107, 711-714.

7. Karas, M., Bachmann, D., Bahr, U. & Hillenkamp, F. (1987) Int. J. Mass Spectrom. Ion Process. 78, 53-68.

8. Brown, R. S. & Lennon, J. J. (1995) Anal. Chem. 67, 1990-1999.

9. Rue, E. & Bruland, K. (2001) Mar. Chem. 76, 127-134.

10. Zafiriou, O. C., Voelker, B. M. & Sedlak, D. L. (1998) J. Phys. Chem. A 102, 5693-5700.

11. Šima, J. & Makáňova, J. (1997) Coordination Chem. Rev. 160, 161-189.

12. Burkholder, J. M., Ramsdell, J. S., Moeller, P. D. R., Gordon, A. S., Lewitus, A. J., Glasgow, H. B., Marshall, H. G. & Morton, S. L. (2004) in Proceedings, Tenth International Conference on Harmful Algae, eds. Steidinger, K. A., Landsberg, J. A., Tomas, C. R. & Vargo, G. A. (Florida Depart. Environ. Conserv., St. Petersburg), in press.

13. Glasgow, H. B., Burkholder, J. M., Mallin, M. A., Deamer-Melia, N. J. & Reed, R. E. (2001) Environ. Health Perspect. 109, 715-730.

14. Humphries, E. M. (2002) Pfiesteria piscicida and Pfiesteria shumwayae: Associations with Environmental Parameters (Delaware Dept. Natural Resources Environ, Control, Dover).

15. Bowers, H. A., Tengs, T., Glasgow, H. B., Burkholder, J. M. & Oldach, D. W. (2000) Appl. Environ. Microbiol. 66, 4641-4648.

16. Burkholder, J. M., Glasgow, H. B. & Deamer-Melia, N. J. (2001) Phycologia 40, 186-204.