ARTICLE pubs.acs.org/est
Harmful Dinoflagellate Ostreopsis cf. ovata Fukuyo: Detection of Ovatoxins in Field Samples and Cell Immunolocalization Using Antipalytoxin Antibodies Giorgio Honsell,† Marco De Bortoli,‡ Sabrina Boscolo,§ Carmela Dell’Aversano,|| Cecilia Battocchi,^ Giampaolo Fontanive,# Antonella Penna,^ Federico Berti,# Silvio Sosa,‡ Takeshi Yasumoto,r Patrizia Ciminiello,|| Mark Poli,O and Aurelia Tubaro*,‡ †
Agriculture and Environmental Sciences Department, University of Udine, Via delle Scienze 91-93, 33100 Udine, Italy Life Sciences, University of Trieste, Via Alfonso Valerio 6, 34127 Trieste, Italy § Life Sciences Department, University of Trieste, Via Licio Giorgieri 10, 34127 Trieste, Italy Chemistry of Natural Products Department, University of Naples, Via Domenico Montesano 49, 80131 Naples, Italy ^ Biomolecular Science Department, University of Urbino, Viale Trieste 296, 61121 Pesaro, Italy # Chemical and Pharmaceutical Science Department, University of Trieste, Via Licio Giorgieri 1, 34127 Trieste, Italy r Japan Food Research Laboratories, Tama Laboratory, 6-11-10 Nagayama, Tama-shi, Tokyo 206-0025, Japan O United States Army Medical Research Institute of Infectious Diseases, Fort Detrick, Maryland 21702-5011, United States
)
‡
bS Supporting Information ABSTRACT: Ostreopsis cf. ovata, a benthic dinoflagellate often blooming along the Mediterranean coasts, has been associated with toxic events ranging from dyspnea to mild dermatitis. In late September 2009, an Ostreopsis cf. ovata bloom occurred in the Gulf of Trieste (Northern Adriatic Sea; Italy), causing pruritus and mild dermatitis in beachgoers. An integrated study was initiated to characterize Ostreopsis cells by light and confocal microscopy, PCR techniques, immunocytochemistry, and high resolution liquid chromatographymass spectrometry (HR LC-MS). The presence of Ostreopsis cf. ovata of the Atlantic/ Mediterranean clade was unambiguously established by morphological and genetic analyses in field samples. Several palytoxin-like compounds (ovatoxin-a,-b,-c,-d,-e) were identified by HR LC-MS, ovatoxin-a being the most abundant (4564 pg/cell). Surprisingly, no palytoxin was detected. For the first time, monoclonal and polyclonal antipalytoxin antibodies revealed the intracellular cytoplasmic localization of ovatoxins, suggesting their cross-reactivity with these antibodies. Since harmful dinoflagellates do not always produce toxins, the immunocytochemical localization of ovatoxins, although qualitative, can provide an early warning for toxic Ostreopsis cells before their massive diffusion and/or concentration in seafood.
’ INTRODUCTION The toxic benthic dinoflagellate genus Ostreopsis Schmidt presents a temperate tropical geographical distribution from about 45N to 45S.13 Ostreopsis ovata Fukuyo was first described in the Ryukyu Islands (Japan),4 and since 1994 was widely distributed also in temperate waters of the Mediterranean Sea.5 Since 1998,6 Ostreopsis has been repeatedly associated with toxic events and reports of its presence have been increasing in many coastal areas.79 Health problems ascribed to aerosols or cutaneous exposure to seawater during Ostreopsis blooms have been recently reported in different areas of the Mediterranean Sea.1014 The most relevant episode occurred along the northwestern coast of Italy, close to Genoa, where in July 2005 more than 200 people were admitted to the local hospitals. r 2011 American Chemical Society
Simultaneously with an Ostreopsis bloom, people staying at the seaside for recreational or work activities developed two or more of these symptoms: dyspnea, rhinorrhea, fever, sore throat, cough, conjunctivitis, headache, nausea/vomiting, and dermatitis.11 Chemical analysis of Ostreopsis net samples collected during the bloom revealed the presence of putative palytoxin and ovatoxin-a, a new palytoxin analogue.15,16 Ostreopsis blooms may pose a serious risk to human health due to the production of palytoxin-like compounds.1722 The presence of putative Received: April 21, 2011 Accepted: July 15, 2011 Revised: July 6, 2011 Published: July 15, 2011 7051
dx.doi.org/10.1021/es201373e | Environ. Sci. Technol. 2011, 45, 7051–7059
Environmental Science & Technology palytoxin and five new palytoxin analogues, named ovatoxin-a, -b, -c, -d, and -e, was recently highlighted in natural populations (ovatoxin-a) and cultured cells (ovatoxin-a, -b, -c, -d, -e) of O. cf. ovata collected along the Italian coasts15,16,23,24 and identified through high resolution liquid chromatographymass spectrometry (HR LC-MS). Also in the Gulf of Trieste (North-Eastern coast of Italy), at the end of September 2009 some bathers experienced itching sensations and mild dermatitis after contact with seawater, although none required hospitalization (Kokelj, personal communication). Concomitantly, an outbreak of Ostreopsis, with millions of cells per liter, occurred in shallow waters, on pebbles and epiphytic on macroalgae. In this study an integrated approach was used to characterize Ostreopsis cells by light and confocal microscopy, PCR techniques, immunocytochemistry, and high resolution liquid chromatographymass spectrometry (HR LC-MS). Antibodies are one of the most common types of probes used in harmful algal bloom research,25 but few studies use antitoxin antibodies directly on dinoflagellate cells. This approach has the advantage of identifying toxin producing cells, while other probes (DNA probes, other immunological probes, lectins) give no specific information on cell toxin content and cannot distinguish toxic from nontoxic strains within the same species.26 Some anti-PLTX antibodies were previously developed,2730 but none have been directly applied to dinoflagellate cells. Our immunocytochemical study was carried out using both monoclonal and polyclonal antipalytoxin antibodies to assess, for the first time, the presence of palytoxin-like compounds inside cells. Results were then confirmed by HR LC-MS.
’ MATERIALS AND METHODS Sampling of Natural Algal Populations. Macroalgae were collected from the upper infralitoral rocks of Aurisina marina (45 440 2500 N, 13 400 0600 E) in the Gulf of Trieste (Italy, Northern Adriatic Sea) on 23rd and 28th September 2009. Seawater and pebbles covered by brown films were sampled in shallow waters (30 to 100 cm) of the adjacent rocky beach of Canovella de’ Zoppoli (45 440 5800 N, 13 390 2000 E) on 28th September 2009. The sample Canovella 1 was collected at the surface 10 m from the beach and the sample Canovella 2 at a depth of 30 cm above pebbles closer to the beach. Each sample was divided into subsamples for the taxonomic identification, immunocytochemistry, and toxin analyses. Taxonomic Identification: Light and Epifluorescence Microscopy. Observations were carried out on unfixed and fixed (1% Lugol acid solution31) samples using a Leitz Diavert inverted microscope (Ernst Leitz Wetzlar; Wetzlar, Germany) using bright field and phase contrast illumination. Macroalgal thalli and seawater samples were first observed unfixed to detect the presence of benthic dinoflagellates. Cell counting in seawater was carried out by the Uterm€ohl method32 at 500x magnification on fixed samples. Cells were observed by epifluorescence microscopy using a Leitz Labovert inverted microscope (Ernst Leitz Wetzlar) at 400x. Observations were carried out using two different filter sets (exciter filter BP450-490, barrier filter LP515; exciter filter BP515-560, barrier filter LP590) and after staining with Calcofluor White M2R (Polysciences; Warrington, USA) with UV excitation (exciter filter BP340-380, barrier filter LP425) to show thecal plates.33
ARTICLE
Table 1. Molecular Formulae (M), Retention Times (Rt), and Most Abundant Ion Peaks of Palytoxin and Ovatoxins Used in Quantitative HR LC-MS Analysesb toxin
M
palytoxin
Rta (min) [M+2HH2O]2+ [M+2H+K]3+
C129H223N3O54
10.78
1331.7436
906.8167
ovatoxin-a C129H223N3O52
11.45
1315.7498
896.1572
ovatoxin-b C131H227N3O53
11.28
1337.7623
910.8318
ovatoxin-c C131H227N3O54
10.90
1345.7584
916.1628
ovatoxin-d C129H223N3O53 and -ea
11.07
1323.7456
901.4884
a
Ovatoxin-d and -e are structural isomers, coeluting under the chromatographic conditions used, that can be distinguished by HR LC-MS/MS analysis.23 b Refer to Methods for stationary and mobile phase conditions.
Taxonomic Identification: Molecular analysis. Ten samples of macrophytes were scraped from the surfaces of rocks and analyzed by molecular PCR assay. Samples were processed for total DNA, and PCR amplification using genus and speciesspecific primers was performed as described in Battocchi et al.34 and Penna et al.35 Toxin Analysis: Extraction. Algal samples were filtered by gravity on 0.45 μm filters (Durapore HV membrane filters, Millipore), and pellets were frozen at 20 C until extraction. Pellets were extracted thrice with 50 mL of methanol/water (1:1v/v) following the procedure previously reported.15,16,23 Crude extracts were analyzed directly by high resolution liquid chromatographymass spectrometry (5 μL injected). Toxin Analysis: High Resolution Liquid ChromatographyMass Spectrometry (HR LC-MS). HR LC-MS was performed on an Agilent 1100 LC binary system (Palo Alto, USA) coupled to a hybrid linear ion trap LTQ Orbitrap XL Fourier Transform MS (FTMS) equipped with an ESI ION MAX source (Thermo-Fisher; San Jose, USA). LC-MS conditions reported previously were used.23 HR full MS experiments (positive ions) were acquired in the range m/z 8001400 at a resolving power of 15,000. Extracted Ion Chromatograms (XIC) were obtained from the HR full MS spectra by selecting the most abundant peaks of the [M+2HH2O]2+ and [M+2H+K]3+ ion clusters of each compound23 (Table 1). A mass tolerance of 5 ppm was used. The chromatographic peaks were identified by comparing their retention times (Table 1) and associated HR full MS spectra (Figure 2) to those of ovatoxins contained in a reference O. ovata extract previously characterized and analyzed under the same experimental conditions.23 Identity of ovatoxins was further confirmed through HR LC-MS/MS experiments (data not shown). Peak areas were measured, and the calibration curve of palytoxin standard (Wako Chemicals; Neuss, Germany) was used in quantitative studies (triplicate injection at four levels of concentrations: 25, 12.5, 6.25, and 3.13 ng/mL). Linearity of the calibration curve was indicated by a correlation coefficient (R2) of 0.9985. Molar responses were assumed to be the same as that of palytoxin, basing on structural similarities between palytoxin and ovatoxins.23 Results were corrected based on recovery efficiency of the extraction procedure (98%).15 Measured limits of quantitation (LOQ) and detection (LOD) for our palytoxin standard were 3.13 and 1.9 ng/mL, respectively. Taking into account the efficiency of the extraction procedure,15 LOQ and LOD of the method were 20 and 13 fg/cell, respectively. 7052
dx.doi.org/10.1021/es201373e |Environ. Sci. Technol. 2011, 45, 7051–7059
Environmental Science & Technology Polyclonal anti-PLTX Antibody Production: Conjugates Preparation. PLTX Activation. 0.125 mg of 4-(N-maleimido-
methyl)-cyclohexane-carboxylic acid N-hydroxy-succinimide ester was dissolved in 0.1 mL of dimethylformamide, and the solution was added to 0.2 mg of PLTX dissolved in 1.7 mL of 0.1 M phosphate buffer, pH 7.5. The mixture was then stirred for 8 h at room temperature (RT). Carrier Protein Activation. Two mg of bovine serum albumine (BSA) was dissolved in 0.8 mL of 25 mM borate buffer pH 9.0, and 0.21 mg of 2-iminothiolane was added. The reaction was incubated for 1 h at RT, and then the buffer was changed to 0.1 M phosphate pH 6.6, containing 1 mM EDTA. The protein was then purified over a short G25 column using the same buffer as eluent. Conjugation. Activated PLTX (50-fold molar excess) was added to the activated carrier protein and stirred for 1 h at RT before dialysis against 500 mL of phosphate buffer (sodium chloride 0.14 M; potassium chloride 2.7 103 M; dibasic sodium phosphate 0.01 M; potassium dihydrogen phosphate 1.76 103 M; pH 7.3). UV analysis was performed at 250 nm (PLTX maximum, BSA minimum) and 280 nm (BSA maximum), and an apparent PLTX:BSA ratio of 34 was estimated. Polyclonal anti-PLTX Antibodies Production: Rabbits Immunization. Two New Zealand White rabbits (2 kg; Harlan Laboratories; San Pietro al Natisone, Italy) were kept at controlled temperature (21 ( 1 C) and humidity (6070%), with a fixed artificial light cycle (7.00 a.m.7.00 p.m.). Animals were caged individually using dust free poplar chips for bedding and fed a standard diet for rabbits (Harlan Laboratories). Animals were immunized intradermally with 200 μg of PLTXBSA conjugate suspended in phosphate buffered saline (PBS) and Complete Freund0 s adjuvant (50:50v/v, 1 mL total volume). After four weeks, the animals were boosted intradermally with PLTX-BSA (100 μg) suspended PBS and incomplete Freund0 s adjuvant (50:50v/v, 1 mL). Subsequent boosts occurred every four weeks by subcutaneous and intramuscular injections of 75 and 25 μg of PLTX-BSA, respectively. Antibody titers were tested by ELISA 10 days after each boost. At the end of the treatment period animals were anaesthetized by intramuscular injection of Zoletil and Xylazine (15 and 5 mg/kg, respectively; Virbac; Milan, Italy) and exanguinated, in accordance with national and international guidelines. Experiments were carried out at the University of Trieste in conformity with Italian D.L. n.116 of 27 January 1992 and associated guidelines in the European Communities Council Directive of 24 November 1986 (86/609 ECC) concerning animal welfare and appendix A of the European Convention ETS123. Monoclonal anti-PLTX Antibody. Mouse monoclonal antiPLTX antibody 73D3 was isolated from hybridoma cultures at the US Army Medical Research Institute of Infectious Diseases, Fort Detrick, Maryland, USA. Immunocytochemistry. Immunocytochemistry was carried out on Ostreopsis cf. ovata and Coolia monotis. Aliquots of cells were collected and pelleted at 300 g for 5 min at RT. Cells were fixed for 24 h with paraformaldehyde (2% in PBS) and sectioned (10 μm) in a cryostat (Leica; Wetzlar, Germany). Sections were collected on polyornithine-coated slides and dried overnight at RT. Nonspecific sites were blocked with 10% normal goat serum in 50 mM Tris/HCl, 0.15 M sodium chloride, 2% BSA, pH 7.5 for 30 min at RT. Sections were then incubated with monoclonal or polyclonal anti-PLTX antibodies overnight at 4 C.
ARTICLE
Preimmune rabbit serum was used as negative control. After four washes with 50 mM Tris/HCl, 0.15 M sodium chloride, pH 7.5 (TBS), cells were incubated for 1 h at RT with Alexa Fluor 488 goat antimouse IgG or Alexa Fluor 488 goat antirabbit IgG (1:200 dilution in TBB) (Invitrogen; Milan, Italy) and washed four times with TBS. Nuclei were stained with Hoechst 33258 (Sigma-Aldrich) (10 ng/μL in PBS) for 5 min at RT. Slides were then washed twice with PBS and with Milli-Q water and mounted with Mowiol 488 mounting medium (Sigma-Aldrich). Evaluation of staining was carried out visually by a double blind procedure with a Nikon E800 fluorescence microscope (Nikon Instruments Europe; Kingston, England) using two different filter sets for Hoechst 33258 (excitation filter: 340380 nm, barrier filter: 435485 nm) and Alexa Fluor 488 (excitation filter: 460500 nm, barrier filter: 510560 nm), respectively. Sections were also observed with a Nikon C1si confocal microscope (Nikon Instruments Europe), using a 60X Plan Apo Oil objective. The system was operated with a pinhole size of one Airy disk. Electronic zoom was kept at minimum values for measurements to reduce potential bleaching. The 488 nm laser line was used to visualize the specific staining of antibody coupled with Alexa Fluor 488. The 561 nm laser line was used to visualize the endogenous autofluorescence due to cellular photosynthetic pigments. A series of optical images at 200 nm increments along the 00 z00 axis of the algae sections were acquired for the entire thickness of the section (10 μm). Images were processed for z-projection and for illustration purposes by using ImageJ (NIH; Bethesda, USA) and Adobe Photoshop CS2 (Adobe Systems; San Jose, USA).
’ RESULTS Late in September 2009, the harmful dinoflagellate Ostreopsis appeared along the rocky coast between Aurisina and the beach of Canovella de’ Zoppoli, where it had been reported in previous years.36 However, the cell counts were much greater in 2009 than previously. By 23rd September several cells were found on the red alga Corallina officinalis L. growing just below the sea level on the rocks outside Aurisina marina. The cells covered the thalli and were attached to a network of tiny filaments formed between the algal branches (Figure 1A). On 28th September a bloom was observed in a small protected pool in shallow waters (20100 cm) near the beach of Canovella de’ Zoppoli. Bottom pebbles were covered by brownish strips, and some were detached and floated in the water. Microscopic observation revealed high numbers of Ostreopsis cells, up to 6 106 cells/L. An itching sensation of the skin was felt during sampling of the shallow water above the pebbles. Similarly, some people experienced an itching sensation and developed mild dermatitis after swimming along the beach. Microscopic analysis of samples collected from the pool revealed high concentrations of benthic dinoflagellates (Ostreopsis cf. ovata with few cells of Coolia monotis Meunier and Prorocentrum sp.), colonial cyanobacteria (mainly Oscillatoriales), and diatoms (Nitzschia cf. reversa W. Smith, some Naviculaceae, Gyrosigma sp., Coscinodiscus sp.). The highest abundance of Ostreopsis cf. ovata was found in very shallow waters in front of the beach (sample Canovella 2 with 6.7 106 cells/L), while the cell concentration was lower near the rocks surrounding the pool (sample Canovella 1 with 9.3 105cells/L). Taxonomic Identification. Ostreopsis cells were ovate and tear-shaped and presented a dorsoventral diameter of 4865 μm and transdiameter of 3146 μm with a max/min diameter ratio 7053
dx.doi.org/10.1021/es201373e |Environ. Sci. Technol. 2011, 45, 7051–7059
Environmental Science & Technology
ARTICLE
Figure 1. Ostreopsis cf. ovata, 28th September 2009, Aurisina (Gulf of Trieste). Light and epifluorescence microscopy. A. O.cf.ovata cells epiphytic on the red alga Corallina officinalis L. Scale bar 100 μm. B. Cells fixed in Lugol’s, phase contrast light microscopy. Scale bar 10 μm. C. Living cells. Epifluorescence microscopy (exciter filter 450490 nm, barrier filter 515 nm). Chloroplasts red chlorophyll autofluorescence is evident. Scale bar 10 μm. D. Epithecal plates seen after Calcofluor White M2R staining by epifluorescence microscopy (exciter filter 340380 nm, barrier filter 425 nm). Scale bar 20 μm. E. Epithecal plate pattern observed by phase contrast after cell squashing. Scale bar 10 μm. F. Hypothecal plate pattern observed by phase contrast after cell squashing. Scale bar 10 μm.
between 1.33 and 1.74 (Figure 1B). Living cells revealed the abundant presence of translucent rounded bodies in the central part of the cell which prevented clear visualization of chloroplasts, visible only for the yellow greenish coloration they give to the cell. Chloroplasts, however, were clearly visible by epifluorescence microscopy (Figure 1C), giving a strong red or orangered emission by blue or green excitation for chlorophyll and carotenoids, respectively. The thecal plate pattern and shape were observed after cell squashing or staining with Calcofluor White M2R (Figure 1D-F) and fitted well with previous Ostreopsis ovata descriptions.1,4 Molecular Analysis. PCR for the presence of Ostreopsis species was carried out on Lugolfixed samples. Amplification confirmed the presence only of the O. cf. ovata genotype in all analyzed samples. Furthermore, final alignment of Ostreopsis sequences with the Genbank sequence database of Ostreopsis species confirmed that the Ostreopsis isolates belonged to the O. cf. ovata Atlantic/Mediterranean clade (see Supporting Information, Figure S1). Toxin Analysis. The O. cf. ovata samples Canovella 1 and 2 were separately extracted. Crude extracts were analyzed by high resolution (HR) LC-MS and compared to both a palytoxin standard and a culture extract of Adriatic O. ovata previously characterized; this latter was used as reference in containing a putative palytoxin and all the ovatoxins so far known (Table 1). 23
While the structure of palytoxin has been fully elucidated17,18 and a standard commercially available, ovatoxins have yet to be isolated and structurally elucidated. The following features of ovatoxin-a, -b, -c, d-, and -e are currently known: i) retention times (Table 1); ii) molecular formulas, determined by the interpretation of monobi- and tricharged ions in their relevant HR full MS spectra; iii) elemental composition of diagnostic fragment ions in their relevant HR MS/MS spectra.23 Figure 2 shows HR full MS spectrum of Canovella 2 sample in comparison to that of palytoxin. Under these conditions, each ovatoxin produces a complex mixture of ionized species, either bi- or tricharged. Mass scale expansion of the most abundant [M +2HH2O]2+ and [M+2H+K]3 ion clusters of each ovatoxin is reported in Figure S2 (Supporting Information). No palytoxin was detected in the analyzed samples (LOD for palytoxin in pellet extract was 13 fg/cell), whereas the presence of all the ovatoxins so far known (ovatoxin-a, -b, -c, -d, and -e)23 was demonstrated. Both retention times and HR full MS data matched those of ovatoxins contained in the reference O. ovata culture extract analyzed under the same experimental conditions.23 Assignment of ion species was performed, and errors in assessment of molecular formulas were all below 2 ppm. HR LC-MS/MS experiments were carried out for further confirmation (data not shown) and compared to data previously reported.23 7054
dx.doi.org/10.1021/es201373e |Environ. Sci. Technol. 2011, 45, 7051–7059
Environmental Science & Technology
ARTICLE
Figure 2. HR full MS spectrum (positive ions) of Canovella 2 sample (all scans in the 10.611.5 min window were summed) (A) versus that of palytoxin (B). The most intense tri- and bicharged ion clusters of each PLTX-like compound are indicated. Ion assignment of the other peaks has been reported previously. 23 Refer to Materials and Methods for LC-MS conditions used.
Extracted ion chromatograms (XICs) of the most abundant ions (Table 1) were used to determine individual and total ovatoxin content of each sample. Figure 3 shows XICs of ovatoxins contained in Canovella 2 sample versus XIC of palytoxin standard (25 ng/mL). Because ovatoxin standards are lacking, the calibration curve of our palytoxin standard was used in quantitative studies (R2 = 0.9985; LOD = 1.9 ng/mL; LOQ = 3.13 ng/mL), and ovatoxins were assumed to present the same molar response as palytoxin. This assumption is supported by structural similarities between palytoxin and ovatoxins, as indicated by small differences in elemental composition (Table 1) and by structural hints provided by HR LC-MS/MS experiments previously reported.23 Table 2 reports toxin content expressed as pg/cell and percentage of the total toxin content. Total toxin contents of the Canovella 1 and 2 samples were 58 and 72 pg/cell, respectively. In both samples, ovatoxin-a was dominant, representing 7789% of the total toxin content. Both Canovella 1 and 2 samples contained ovatoxin-b (14% and 5%, respectively) as well as ovatoxin-d/e (9% and 4%, respectively). Ovatoxin-c was contained only in the Canovella 2 sample, whereas it was not detected (
