<%BANNER%>

Paralytic Shellfish Toxins Characterized from Lyngbya Wollei Dominated Mats Collected from Two Florida Springs

Permanent Link: http://ufdc.ufl.edu/UFE0042926/00001

Material Information

Title: Paralytic Shellfish Toxins Characterized from Lyngbya Wollei Dominated Mats Collected from Two Florida Springs
Physical Description: 1 online resource (97 p.)
Language: english
Creator: FOSS,AMANDA J
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre: Fisheries and Aquatic Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Lyngbya wollei is a commonly observed cyanobacterium in Florida?s spring fed systems and is considered a nuisance organism by forming large benthic and floating mats. Two Florida Springs, Silver Glen Springs, and Blue Hole Spring (Ichetucknee) with standing crops of Lyngbya were sampled and mats characterized via microscopy. Molecular data was acquired by the amplification and sequencing of stxA and stxG genes and a section of 16S rRNA from a filament collected from Silver Glen Natural Well. Paralytic shellfish toxins (PSTs) were characterized utilizing High Performance Liquid Chromatography (HPLC) coupled with Fluorescence (FL) and HPLC coupled with Mass Spectrometry (MS). Extraction techniques may convert less toxic L. wollei toxins to more toxic decarbamoylgonyautoxins (dcGTX2&3) and decarbamoylsaxitoxin (dcSTX). Anything stronger than 0.01 M HCl should not be utilized if preservation of the original toxin profile is desired. It was also determined that the use of Solid Phase Extraction did not provide additional support when analyzing the toxins with LC/MS or LC/FL. Pre-column oxidation and LC fluorescence was not a sufficient method of determining PST variants extracted from L. wollei mats since oxidation products are either converted or co-elute with other known PSTs and do not allow for characterization. LC/MS is the only method explored in this study with the ability to do so, but without suitable standards for LC/MS/MS, quanititation and full confirmation could not be made for all the L. wollei Toxins. Analysis with LC/MS/(MS) of extracted algal material did show that dcGTX23 and dcSTX were present in L. wollei mats collected in Florida springs, with evidence pointing to the presence of all L. wollei toxins. In future work with L. wollei derived PSTs, standards of the L. wollei toxins 1-6 should be utilized for LC/MS/(MS) or post column LC/FL quantification to accurately determine total PST content.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by AMANDA J FOSS.
Thesis: Thesis (M.S.)--University of Florida, 2011.
Local: Adviser: Phlips, Edward J.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042926:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042926/00001

Material Information

Title: Paralytic Shellfish Toxins Characterized from Lyngbya Wollei Dominated Mats Collected from Two Florida Springs
Physical Description: 1 online resource (97 p.)
Language: english
Creator: FOSS,AMANDA J
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre: Fisheries and Aquatic Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Lyngbya wollei is a commonly observed cyanobacterium in Florida?s spring fed systems and is considered a nuisance organism by forming large benthic and floating mats. Two Florida Springs, Silver Glen Springs, and Blue Hole Spring (Ichetucknee) with standing crops of Lyngbya were sampled and mats characterized via microscopy. Molecular data was acquired by the amplification and sequencing of stxA and stxG genes and a section of 16S rRNA from a filament collected from Silver Glen Natural Well. Paralytic shellfish toxins (PSTs) were characterized utilizing High Performance Liquid Chromatography (HPLC) coupled with Fluorescence (FL) and HPLC coupled with Mass Spectrometry (MS). Extraction techniques may convert less toxic L. wollei toxins to more toxic decarbamoylgonyautoxins (dcGTX2&3) and decarbamoylsaxitoxin (dcSTX). Anything stronger than 0.01 M HCl should not be utilized if preservation of the original toxin profile is desired. It was also determined that the use of Solid Phase Extraction did not provide additional support when analyzing the toxins with LC/MS or LC/FL. Pre-column oxidation and LC fluorescence was not a sufficient method of determining PST variants extracted from L. wollei mats since oxidation products are either converted or co-elute with other known PSTs and do not allow for characterization. LC/MS is the only method explored in this study with the ability to do so, but without suitable standards for LC/MS/MS, quanititation and full confirmation could not be made for all the L. wollei Toxins. Analysis with LC/MS/(MS) of extracted algal material did show that dcGTX23 and dcSTX were present in L. wollei mats collected in Florida springs, with evidence pointing to the presence of all L. wollei toxins. In future work with L. wollei derived PSTs, standards of the L. wollei toxins 1-6 should be utilized for LC/MS/(MS) or post column LC/FL quantification to accurately determine total PST content.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by AMANDA J FOSS.
Thesis: Thesis (M.S.)--University of Florida, 2011.
Local: Adviser: Phlips, Edward J.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042926:00001


This item has the following downloads:


Full Text

PAGE 1

1 PARALYTIC SHELLFISH TOXINS CHARACTERIZED FROM LYNGBYA WOLLEI DOMINATED MATS COLLECTED FROM TWO FLORIDA SPRINGS By AMANDA FOSS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 AMANDA FOSS

PAGE 3

3 To my wonderful parents

PAGE 4

4 ACKNOWLEDGMENTS I appreciate the support provided me by GreenWater Laboratories, Mark Aubel for his wonderful direction and aid with toxin analysis and Andrew Chapman for his expertise in phycology and sampling I thank the chair Ed Phlips, and members of my committee Nancy Szabo and Karl Havens, for their priceless advice and keen research assistance. I thank Mete Yilmaz for his work on the molecular portion of th is study, as well as expertise I thank Andrew Reich with t he Florida Department of Health for collaboration with GreenWater Labs and the research that kick started this project and, o f course a big thank you to Alicia Plakotaris and Johnny May for their help in the field.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 Objectives ................................ ................................ ................................ ............... 12 Background ................................ ................................ ................................ ............. 13 Lyngbya Growth in Florida Springs ................................ ................................ .. 13 Lyngbya Toxin Production ................................ ................................ ................ 14 Paralytic Shellfish Toxins ................................ ................................ .................. 15 Properties and toxicity ................................ ................................ ................ 15 Potential risks and trophic transfer ................................ ............................. 16 Toxin Extraction ................................ ................................ ................................ 17 Toxin An alysis Techniques ................................ ................................ ............... 18 Molecular Biology of Lyngbya and Saxitoxin gene ................................ ........... 19 2 MATERIALS AND METHODS ................................ ................................ ................ 21 Sampling ................................ ................................ ................................ ................. 21 Site Description ................................ ................................ ................................ 21 Collection ................................ ................................ ................................ .......... 22 Standards and Reagents ................................ ................................ ........................ 24 Standards for Instrument Calibration and Quantification ................................ .. 24 Reagents ................................ ................................ ................................ .......... 24 Sample Preparation ................................ ................................ ................................ 25 Qualitative Analysis of Filamentous Macroalgae Samples ............................... 25 DNA Isolation, Polymerase Chain Reaction, and Sequencing ......................... 25 Toxin Samples ................................ ................................ ................................ .. 27 Paralytic Shellfish Toxin Extraction ................................ ................................ ......... 28 Extraction Assessment ................................ ................................ ..................... 28 SPE and Lyophilization Assessment ................................ ................................ 28 Final Extra ction Protocol ................................ ................................ ................... 29 Paralytic Shellfish Toxin Analysis ................................ ................................ ........... 29 HPLC/ Fluorescence ................................ ................................ ........................ 29 Peroxide Oxidation ................................ ................................ ..................... 29 Periodate Oxidation ................................ ................................ ................... 30 Chromatographic and Fluorescence Conditions ................................ ........ 31

PAGE 6

6 Detection Limits ................................ ................................ ......................... 32 Sample Analysis ................................ ................................ ........................ 32 HPLC/MS/(MS) ................................ ................................ ................................ 32 Chromatographic and Mass Spectrometry Conditions ............................... 32 Detection Limits ................................ ................................ ......................... 33 Sample Analysis ................................ ................................ ........................ 34 3 RESULTS ................................ ................................ ................................ ............... 48 Field Data ................................ ................................ ................................ ............... 48 Water Quality ................................ ................................ ................................ .... 48 Field Observations ................................ ................................ ........................... 50 Algal Analysis ................................ ................................ ................................ ......... 50 Molecular Biology ................................ ................................ ................................ ... 51 HPLC/Fluorescence ................................ ................................ ................................ 52 Peroxide Oxidation Products ................................ ................................ ............ 52 Extraction Assessment ................................ ................................ ..................... 53 SPE and Lyophilization Assessment ................................ ................................ 55 Sample Analysis ................................ ................................ ............................... 56 HPLC/MS ................................ ................................ ................................ ................ 56 Extraction Assessment ................................ ................................ ..................... 56 SPE and Lyophilization Assessment ................................ ................................ 57 Sample Analys is ................................ ................................ ............................... 57 4 DISCUSSION ................................ ................................ ................................ ......... 87 Toxin Extraction ................................ ................................ ................................ ...... 87 Analytical Methods ................................ ................................ ................................ .. 88 Toxin Risk Assessment ................................ ................................ ........................... 89 LIST OF REFERENCES ................................ ................................ ............................... 94 BIOGRAPHICAL SKET CH ................................ ................................ ............................ 97

PAGE 7

7 LIST OF TABLE S Table page 2 1 List of sample collections ................................ ................................ .................... 42 2 2 Saxitoxin stru ctural variants ................................ ................................ ................ 43 2 3 LC/FL gradient ................................ ................................ ................................ .... 45 2 4 MDLs for PSTs via fluorescence in elution order ................................ ................ 45 2 5 LC/MS gradient ................................ ................................ ................................ ... 45 2 6 MDLs and retention times of PSTs analyzed via LC/MS/(MS) ........................... 46 2 7 Molecu lar and fragmentation ions monitored LC/MS ................................ .......... 47 3 1 Water quality and sample collection data for sample sites ................................ 61 3 2 Water quality data for Blue Hole Spring and Silver Glen Springs from FDEP Initiative Report 2011 ................................ ................................ .......................... 62 3 3 Mat thickness measured at Silver Glen Springs ................................ ................. 63 3 4 Qualitative algae list from all samples ................................ ................................ 64 3 5 pH of each extraction solution before and after boiling ................................ ....... 69 3 6 De carbamoylgonyautoxin 2&3 (dcGTX2&3) concentrations ( g g 1 ) for all sites and sampling months ................................ ................................ ................. 76 3 7 Decarbamoylsaxitoxin (dcSTX) concentrations ( g g 1 ) for all sites and sampling months ................................ ................................ ................................ 77 4 1 Toxin content of L.wollei collected from Alabama and Florida ............................ 93 4 2 Toxicity of PST variants reported in MU/mol and relative toxicity to STX (STX eq) ................................ ................................ ................................ ............. 93

PAGE 8

8 LIST OF FIGURE S Figure page 2 1 Silver Glen Springs Recreational Park photo ................................ ..................... 36 2 2 Series of Silver Glen Main Stem photos ................................ ............................. 37 2 3 Series of Silver Glen Natural Well (Southeastern Vent) photos .......................... 39 2 4 Series of Blue Hole (Jug) Spring photos ................................ ............................. 41 2 5 Conversion of saxitoxin to fluorescent derivative for HPLC/FL detection ........... 44 3 1 Series of micrographs of Lyngbya wollei ................................ ............................ 66 3 2 Series of micrographs of Lyngbya wollei epiphytes, all at 400x .......................... 67 3 3 Fluorescence chromatogram of standard solution mixture of dcGTX &3, dcSTX, GTX 2&3, GTX 5, and STX ................................ ................................ .... 69 3 4 Fluorescence chromatogram of oxidized extraction (0.1 M acetic acid) with peaks appe aring at retention times the same as dcGTX2&3 and dcSTX/dcNEO ................................ ................................ ................................ .... 70 3 5 Extraction technique data (HPLC/FL). Data is represented by normalizing peaks to standards of dcGTX2&3 and dcSTX ................................ .................... 71 3 6 Fluorescence chromatogram of oxidized extraction (0.1 M acetic acid) with peaks appearing at retention times the same as dcGTX2&3 and dcSTX and with 10g STX spiked pre extraction (Blue Hole col lected 8/27/09) ................... 72 3 7 Extraction technique data (HPLC/MS). Includes L. wollei toxins, dcGTX2, dcGTX3 and dcSTX relative MS responses ................................ ....................... 73 3 8 HPLC/MS chromatograms of dcGTX2 & dcGTX3 SIM ions from Blue Hole Spring extracted with two different acids ................................ ............................ 74 3 9 HPLC/MS/MS chromatograms of dcSTX from Blue Hole Spring extracte d with two different acids. ................................ ................................ ...................... 75 3 10 Sum of dcGTX2 and dcGTX3 concentrations for all sites over sampling period. ................................ ................................ ................................ ................ 76 3 11 Concentr ations of dcSTX for all sites over sampling period. .............................. 77 3 12 HPLC/MS SCAN of Silver Glen Springs Main Stem with SIM ions for L. wollei Toxins 1 6 ................................ ................................ ................................ ........... 78

PAGE 9

9 3 13 HPLC/MS SCAN of Silver Glen Springs Natural Well with SIM ions for L. wollei Toxins 1 6 ................................ ................................ ................................ 79 3 14 HPLC/MS SCAN of Blue Hole with SIM ions for L. wollei Toxins 1 6 ................. 80 3 15 Averaged MS responses over period of study for L. wollei Toxins 1 6 at each site ................................ ................................ ................................ ...................... 81 3 16 MS Responses to Lyngbya wollei Toxin #1 detected from all s ites over the sampling period. ................................ ................................ ................................ 8 2 3 17 MS Responses to Lyngbya wollei Toxin #2&3 detected from all sites over the sampling period.. ................................ ................................ ................................ 83 3 18 MS Responses to Lyngbya wollei Toxin #4 detected from all sites over the sampling period. ................................ ................................ .............................. 84 3 19 MS Responses to Lyngbya wollei Toxin #5 detected from al l sites over the sampli ng period. ................................ ................................ ................................ 85 3 20 MS Responses to Lyngbya wollei Toxin #6 detected from all s ites over the sampling period. ................................ ................................ ................................ 86

PAGE 10

10 Abstract of Thesis P resented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of S cience PARALYTIC SHELLFISH TOXINS CHARACTERIZED FROM LYNGBYA WOLLEI DOMINATED MATS COLLECTED FROM TWO FLOR IDA SPRINGS By Amanda Foss May 2011 Chair: Ed ward Phlips Major: Fisheries and Aquatic Sciences Lyngbya wollei is a commonly observed spring fed systems and is considered a nuisance organism by forming large benthic and float ing mats Two Florida s prings Silver Glen Springs, and Blue Hole S pring (Ichetucknee) with standing crops of Lyngbya were sampled and mats characterized via microscopy Molecular data was acquired by the amplification and sequencing o f sxt A and sxt G gen es and a section of 16S r RNA from a filament collected from Silver Glen Natural Well Paralytic shellfish toxins (PSTs) were characterized utilizing High Performance Liquid Chromatography (HPLC) coupled with F luorescence (FL) and HPLC coupled with M ass S p ectrometry (MS). E xtraction techniques may convert less toxic L. wollei toxins to more toxic decarbamoylgonyautoxins (dcGTX2&3) and decarbamoyl saxitoxin (dcSTX) Anything stronger than 0.01 M HCl should not be utilized if preservation of the original tox in profile is desired. It was also determined that the use of Solid Phase E xtraction did not provide additional support when analyzing the toxins with LC/MS or LC/FL. P re column oxidation and LC fluorescence was not a sufficient method of determining PST variants extracted from L. wollei mats since oxidation products are either converted or co elute with other known PSTs and do not allow for

PAGE 11

11 characterization. LC/MS is the only method explore d in this study with the ability to do so but without sui t able standards for LC/MS/MS, quanititation and full confirmation could not be made for all the L. wollei Toxins Analysis with LC/MS/(MS) of extracted algal material did show that dcGTX23 and dcSTX were present in L. wollei mats collected in Florida springs, w ith evidence pointing to the presence of all L. wollei toxins. In future work with L. wollei derived PSTs, standards of the L. wollei toxins 1 6 should be utilized for LC/MS/ ( MS ) or post column LC/FL quantification to accurately determine total PST conten t.

PAGE 12

12 CHAPTER 1 INTRODUCTION Objectives Cyanobacteria (a.k.a blue green algae) produce a m yriad of active metabolites which are known to induce toxic responses, with a long history of directly harming human and animal populations around the world (Sivone n and Jones, 1999). In Lyngbya wollei which form s large benthic and floating mats L. wollei has long been suspected of producing a myriad of toxins ( Seifert et al. 2007 Berry et al. 200 4 Teneva et al. 2002 Carmichale et al. 1997). Due to its widespread presence in many of Florida systems used for human recreation, it has become a management concern ( Cowell and Botts 1994 Stevenson et al. 2007 ) T oxins of interest include dermatoxi ns (toxins that damage the skin), hepatotoxins (toxins that damage the liver), and neurotoxins (toxins that damage nerve cells). Anecdotal reports of adverse skin reactions (rashes, hives, and blisters), gastrointestinal disorders, respiratory illness, an d even temporary loss of consciousness following potential exposure to cyanobacteria in Florida waterways has been documented by the Florida Department of Health (FDOH) ( personal communication) but specific studies relating to toxin s are limited The F DOH initiated an evaluation of freshwater Lyngbya and its toxins in Florida s prings in 2004 (Miller 2006) The study reported the presence which were designated unknown paralytic shellfish toxins (PSTs) The presence of sax itoxin was not confirmed with the use of LC/MS /MS and other PSTs were not analyzed Since the morphologic and molecular identification of L wollei is under some scrutiny (Joyner et al. 2008) and toxin

PAGE 13

13 production ca nnot be determined from simply determin ing the presence of Lyngbya characterizing PST toxin s present in Florida L. wollei mats is an important g oal Based on initial assessment of data, it is hypothesized that L. wollei mats in Florida springs produce Paralytic Shellfish Toxins (PSTs) whic h are not accurately characterized by frequently utilized methods of detection. The goal of this study was to test for the presence of PSTs in L. wollei dominated mat samples from two Florida spring fed pools. Currently, many methods utilized for PST det ection remain largely non specific and present problems when analyzing a complex toxin profile such as the one for PSTs extracted from L. wollei Additionally, improper extraction of the PSTs can inadvertently change the original toxin profile, resulting i n over or under estimation of actual toxicity, so a careful examination of extraction protocols must be conducted to prevent toxin profile shifts ( Oshima 1995, Indresena & Gill 2000 Vale et al. 2008) Background Lyngbya Growth in Florida Springs Macro algae growth in has been on the increase over the last 50 years (Stevenson et al. 2007). Blooms of filamentous cyanobacteria, such as Lyngbya wollei ( Farlow ex Gomont) Speziale and Dyck, have become increasingly responsible for declining habitat and water quality (Cowell and Botts 1994). Recently, a survey submitted to the Florida Department of Environmental Protection (FDEP) evaluating algal growth and nutrients in 21 different springs found that Lyngbya wollei and Vaucheria spp. were the most common macroalgae present (Stevenson et al. 2007) Although Lyngbya abundance was not related to spring water nutrient concentrations, it was positively correlated to human activities around sampling sites and to other indicators of nutrient sup ply (i.e. sediment phosphorus) (Stevenson et al. 2007). It is

PAGE 14

14 even proposed that these increases in macroalgae abundance are in response to other factors, such as lower dissolved oxygen levels that effectively reduce grazer interactions (Heffernamn et al. 2010). Lyngbya Toxin Production Lyngbya (Family Oscillatoriales) is a known toxin producer, with more research conducted on marine species, such as Lyngbya majuscula Although toxin production has not been as thoroughly investigated in freshwater Lyngb ya as in its marine counterpart some information is available in the scientific literature. In Australia, Lyngbya wollei has been shown to produce cylindrospermopsin (CYN) and deoxy cylindrospermopsin, a known hepatotoxin (Seifert et al. 2007). Cylindro spermopsin is toxic to both the liver and kidneys and has the ability to cause severe gastroenteritis. A freshwater species of Lyngbya isolated from the Florida Everglades was shown to inhibits microbial and green algal growth (Berry et al. 2004). Berry et al. (2004) suggested that although extracellular concentrations of this metabolite were low, pahayokolide A may function as an allelochemical, inhibiting the growth of other cyanobact eria or algae that may potentially compete for nutrients or space. Lyngbya aerugineo coerulea, another freshwater species, is common in southern Europe and is known to be toxic to mammals and fish cells (Teneva et al. 2002). Because the chemicals from th is Lyngbya species were crudely extracted, it is unknown which chemical isolate(s) produce the cytotoxic response. S amples of L. wollei collected from Lake Guntersville, Alabama have yielded PSTs that resulted in positive neurotoxic response s in mouse bioa ssay s. The presence of the PSTs decarbamoylgonyautoxins 2&3, decarbamoylsaxitoxin, as well as 6 novel PSTs

PAGE 15

15 ( L. wollei Toxins 1 6) were detected utilizing post column fluorescence deriva tization HPLC techniques, and confirmed with NMR (Carmichael et al. 19 97, Yin et al. 1997, Onodera et al. 1997). This, however, has not been documented in L. wollei samples collected from any other source. T he PSTs produced by the Alabama Lyngbya wollei are less toxic derivatives of PSTs as determined by mouse bioassay. Y in et al. (1997) found that PST production and dry weight increased with calcium concentrations in Lyngbya mats, a phenomenon which has been observed in previous growth studies (Cowell and Botts, 1994). Paralytic Shellfish Toxins Properties and t oxicit y Paralytic Shellfish T oxins (PSTs) are commonly referred to as saxitoxins (or paralytic shellfish poisons). PSTs are a group of neurotoxic alkaloids produced primarily by freshwater cyanobacteria and marine dinoflagellates with over 57 variants reported to date (Wiese et al. 2010) Exposure to these toxins may result in illness, paralysis, and even death. The PSTs reversibly bind to voltage gated sodium channels, inducing neurological symptoms. The degree of the effect varies with the amount of toxin exposed to, the exposure route and which of the variants are present. The different PSTs are grouped by hydrophilicity and by the nature of the side chains (carbamate, acetate, sulfate, etc.). Not all variants are of comparable toxicity and identificatio n of specific PSTs is of importance, not only when analyzing for the toxins, but also when calculating potential risk. When analyzing for PSTs in shellfish for human consumption, a limit of 80 g STX equivalents per 100g of shellfish meat is set as the ma ximum accep t able limit in most countries, with few countries having lower established limits (Campas et al. 2007). For drinking water, there are even fewer

PAGE 16

16 specific guidelines. A level of three g STX eq uivalents L 1 is recommended as the health alert le vel for drinking water in Australia based on documented human toxicity data (Fitzgerald et a l 1999) are determined from bioassay response, rather than analytical varian ts detected. C urrent methods are moving away from the mouse bioass ay approach and toward analytical techniques, so t he European Commission of Marine Biotoxins in Shellfish rate the toxicity of some of the PST s with toxicity equivalency factors (TEF) based on acute toxicity in mice. With saxitoxin representing the most t oxic and rated as 1, the fo llowing toxins are rated as : NEO =1, GTX1 = 1, GTX2 = 0.4, GTX3 = 0.6, GTX4 = 0.7, GTX5 = 0.1, GTX6 = 0.1, C2 = 0.1, C4= 0.1, dc STX = 1, dc NEO = 0.4 dc GTX2 = 0.2, dc GTX3 = 0.4, and 11 hydroxy STX =0.3 (Scientific Opinion of the P anel on Contaminants in the Food Chain on a request from the European Commission on Marine Biotoxins in Shellfish Saxitoxin Group 2009). Unfortunately, this rating effort does not include all the potential variants and scientific literature is not alway s in agreement with regard to toxicity. Potential risks and trophic t ransfer Risk to animals and humans associated with Lyngbya PST production is currently unclear as reports of exposure are not be always linked to toxin production. Direct contact with t he mats may cause mild neurotoxic symptoms in mammals (e.g., tingling, numbne ss, etc.), but ingestion ( direct or indirect) may be required for strong adverse toxic effects. Direct ingestion has not been reported, but indirect ingestion of organisms expose d to PSTs may occur. Currently, it is unknown to what extent organisms feed on Lyngbya or if trophic transfers take place, but some work has been conducted on herbivory Camacho and Thacker (2006) investigated amphipod herbivory on PST producing Lyngbya wollei and found that the presence of PSTs stimulated the feeding of

PAGE 17

17 Hyalella azteca (an omnivorous amphipod), even though the sheath prevented feeding. Their results suggest that Hyalella can tolerate or detoxify the PSTs produced by Lyngbya Whether o r not trophic transfer of PSTs produced by Lyngbya occurs is not yet known. Toxin Extraction Currently, there is not an effective published method for the extraction of PSTs from algal mats such as the ones produced by Lyngbya wollei Extraction methods h ave been published for PSTs from phytoplankton (Ravn et al. 1997), but certain characteristics of Lyngbya (i.e., the presence of accessory pigments and a tough sheath) make extraction difficult. A reproducible extraction method for algal mat material such as Lyngbya is necessary for accurate qualitative and quantitative analysis of PST toxins. Sample collection, homogenization and preparation techniques, extraction of lyophilized algal material and additional clean up techniques such as Solid Phase Extrac tion (SPE) must be considered to improve reproducibility and to reduce interferences and false peaks. Traditional extraction methods (e.g. AOAC Official Method 2005.06), based on human consumption of shellfish, utilize an acidic solution as the extraction media. PSTs are readily soluble in water and are considered heat s Table but they are reported to degrade at higher pH (Botana 2008). Actually, both chemical and biological interconversion of PST variants is a common occurrence, making qualitative and qu antitative determinations difficult. For instance, it has been demonstrated that extraction with 0.1 M hydrochloric acid (via boiling) may convert less toxic variants of PSTs, such as GTX5 to more toxic ones, such as saxitoxin ( STX) (Oshima 1995, Vale et al. 2008, Indresena & Gill 2000). Keeping the boiling times short and reducing the

PAGE 18

18 concentration of acid may reduce the risk of hydrolyzing the less toxic variants to the more toxic ones while keeping the sample at a sufficient low pH to prevent degradat ion and other conversions. In addition, separation by SPE may also aid in reducing/removing many of the chemically active components that could change the extracted PSTs post extraction, thereby allowing for longer hold times of samples prior to analysis. Toxin Analysis Techniques Analysis of extracted algal material has been conducted via multiple methods, including Mouse Bioassay (MBA), Enzyme Linked Immunosorbent Assay (ELISA), receptor binding assays, and Liquid Chromatography coupled with either Mass Spectrometry or Fluorescence (LC/MS or LC/FL). Because MBA is considered unethical, assay approaches for toxins have moved toward analytical/instrumental testing. While ELISA is reportedly highly specific and sensitive, complex biological matrices are kn own to confound and interfere with results. Additionally, ELISA kits for saxitoxins are designed for the assay of only saxitoxin. Because other variants are not reported to have a 1:1 cross reactivity, STX specific ELISAs are not useful quantitatively an d may, depending on what toxin variants are present, severely underestimate the actual concentrations of STX related toxins that are present. Although receptor binding assays are a better measure of how the extract may impact sodium channel receptors, the se assays do not identify the variants present and, unfortunately, the cross reactivity of other similarly acting chemicals may cause interference. Methods employing chromatography include post and pre column oxidation techniques since PSTs have only a w eak natural chromophore for spectrophotometric detection and require s oxidation to detect the compound. Post column oxidation requires a complex setup and

PAGE 19

19 daily maintenance requirements. The Association for Analytical Chemist Official method for shellfis h toxins adopted the pre column oxidation with hydrogen peroxide and periodic acid solutions which is considered by regulators to be the accepted method for the analysis of PSTs. Prechromatographic oxidation was utilized for detection of PSTs present in L yngbya dominated mats in this study. Confirmation of PST toxins and scans for multiple PSTs was conducted using Liquid Chromatography coupled w ith Mass Spectrometry (LC/MS) Molecular Biology of Lyngbya and Saxitoxin gene Distantly related organisms, the dinoflagellates and cyanobacteria, both utilize the same biosynthesis pathway for PST production (Kellman et al. 2008) The saxitoxin biosynthetic gene cluster (sxt) was recently identified in several cyanobacteria species including Cylindrospermopsis rac iborskii T3 (Kellmann et al 2008 ), Anabaena circinalis AWQC131C, Aphanizomenon sp. NH 5 (Mihali et al. 2009) and Raphidiopsis brookii D9 (Stucken et al. 2010). While C. raciborksii sxt biosynthesis gene cluster covers a 35 kb (kilobase pair ) region and c ontains 31 Open Reading Frames ( ORFs ) A. circinalis has 22 ORFs spanning a 29 kb region, Aphanizomenon sp. has 24 ORFs spanning a 27.5 kb region, and R. brookii has 24 ORFs in a 25.7 kb genomic r egion. Kellmann et al (2008 ) proposed that 11 to 14 of the se genes are necessary for the production of the parent saxitoxin molecule and other genes in the biosynthetic cluster encodes for enzymes for the modification of the parent molecule. Therefore differences in the number of encoded enzymes in the sxt biosy nthesis cluster from different organisms p robably reflect the types of PST variants produced by th ese species. Interestingly, sxt X, which is present in C. raciborskii T3, Aphanizomenon flos aquae NH 5, and Lyngbya wollei but absent in A. circinalis AWQC 131C encodes an enzyme similar to

PAGE 20

20 cephalosporin hydroxylase which is proposed to perform N 1 hydroxlation of saxitoxin Indeed while the former two s pecies form N 1 hydroxylated PSTs (e.g. neosaxsitoxin), A circinalis does not produce it and it has yet to be detected in L. w ollei extracted samples (Kellmann et al. 200 8 ). SxtI, a gene proposed to encode for a carbamoyltransferase was naturally mutated in L. wollei previously analyzed from Alabama and encoded an inactive enzy me (Kellmann et al. 200 8). A ccording to Onodera et al. (1997), L. wollei species d o not produce carbamoylated PSTs and it is suggested that this is due to the mutated s x t I gene (Kellmann et al. 2008). PSTs produced by L. wollei may play an important role in the proliferation of the c yanobacterium by influencing grazer interactions. A recent paper characterizing the L. wollei genetics found genes coding for Multi Antimicrobial E xtrusion (MATE) proteins of the NorM family and contains a gene, sxtPER which is similar to permeases of th e Drug and Metabolite Transporter (DMT) family (Mihali et al. 2011). These proteins may increase extracellular transport of PSTs produced by L. wollei which could change our understanding of the role PST production of L. wollei plays but further studies to determine if active transport is occurring would be required.

PAGE 21

21 CHAPTER 2 MATERIALS AND METHOD S Sampling Site Description Three sites were chosen for Lyngbya collection, one in Blue Hole Spring (Ichetucknee State Park) and two from Silver Glen Springs Recreation Area (Ocala National forest). These springs were chosen as they have standing Lyngbya wollei dominated mats observable year round. Silver Glen Springs is located in the Ocala National Forest in Marion County Florida. It is utilized as a recr eation area mainta ined by the United States Department of Agriculture (USDA) Forest Service and has a large combined spring pool emanating second largest freshwater lake. Lake George flows into the St. Johns River out towards the Atlantic. Both freshwater and saltwater species of fish are prominent in both vents and throughout the spring run as it is influenced by natural sources of saline water. The main pool is sectione d by three ropes, two encompass a large Lyngbya mat, meant to keep boaters from entering the spring pool, and the third restricts access to swimmers from the southwest vent. The eastern spring vent is a 1 st magnitude (> 1 00 cubic feet per second (cfs) spr ing shaped as a conical depression, approximately 20 f ee t (ft) deep and is frequented by swimmers. Very little Lyngbya wa s observed in this open section, with pristine waters and a sandy bottom. The southwest vent (also known 2 nd magnitude spring (>10 100 cfs). The spring is a vertical vent, approximately 40 ft deep and surrounded by Lyngbya mats. Both sectioned off areas of the spring were sampled over the period of this study.

PAGE 22

22 Blue Hole Spring (aka Jug Spring) is a 1 st ma gnitude spring in the Ichetucknee state park located in Columbia County FL. The spring directly contributes to the flow of the Ichetucknee River, which flows six miles until emptying into the Santa Fe River. Although recreational activities such as tubin g and canoeing are allowed down the main spring run, the activities in Blue Hole are limited to swimming, cave diving and snorkeling. Blue Hole Spring is one of many that feed into the Ichetucknee system, the second downstream of seven other named springs but with the greatest contributing flow at over 100 cfs. Lyngbya mats in this spring are located along the bottom as well as vertically distributed along the northern face of the spring hole, loosely attached to the roots of the Bald Cypress ( Taxodium d istichum ) trees pictured in Figure 2 4 The v ertically distributed mat wa s the one sampled for this study. Collection The goal of this study was to collect representative L. wollei samples f rom the top of the mats f or toxin analysis A single line transe ct was utilized at each site to assure all collections were made from an area there was L. wollei material. A transect w as established by running survey tape from one end of the mat to the other, approximately down the middle. Transects were re establish ed each collection period in the same direction and approximate area. The largest mat, at Silver Glen Main Stem area, required the longest transect of thirty seven meter s. The tape was run along the bottom across the length of the mat from east to west. The Natural Well transect was nine meters in length, running along the bottom across the mat from north to south. A 5 meter transect was utilized at Blue Hole spring, set up at 1 meter in depth The mat at Blue Hole Spring was not horizontally distribut ed like the mats in Silver Glen Springs.

PAGE 23

23 The Lyngbya grows vertically, loosely attached to the roots of a cypress tree (Figure 2 4; D, E, & F), therefore c ollections were taken at half the max depth of one meter. Algal mat samples were collected as grab samples along the single line transect at each site The five grab collection points were chosen based on random numbers generated in Microsoft E xcel. Each grab was approximately 20g wet and composited invertebrates and debris were rinsed from the composited mat lightly with spring water. The samples were maintained below 10C after collection and for transport. All sample preparation for lyophilization, DNA isolation, and algal analysis was conducted within 4 hours of collection. There were six collection event s where all three sites were visited. One composite d collection per site was taken during a sampling event. A duplicate field collection was collected from one of the sites during each samplin g event. This resulted in 4 total samples collected each event. A total of 24 sample s were collected for this study ( Table 2 1 ). Data acquired from sampling sites included canopy coverage (densiometer, Wildlife Supply Company, Yulee, FL), depth (max, se cchi & sampling), pH, Dissolved Oxygen (DO), specific conductivity, salinity, temperature (YSI MP6600 sonde, YSI Inc., Yellow Springs, OH) and other observations regarding mat size/distribution. Densiometer reading s for overstory density (canopy coverage) we re taken at every sample grab collection ( 5 readings per composited collection ) and averaged for each sample site. Overstory density was determined from raw data by the following calculation: 100% (# unfilled squares x 4.17) = average overstory densi ty.

PAGE 24

24 Standards and Reagent s Standards for Instrument Calibration and Quantification Standards for analysis and instrument calibration were purchased from two sources. The Certified Reference Materials Program of the National Research Council of Canada pr ovided Saxitoxin (STX), Neosaxitoxin (NEO), Decarbamoylneosaxitoxin (dcNEO), Gonyautoxin 1&4 (GTX1&4), Gonyautoxin 2&3 (GTX2&3), Gonyautoxin 5 (GTX5), Decabamoyl Gonyautoxin 2&3 (dcGTX2&3) and N sulfocarbamoyl gonyautoxin 2 and 3 (C1&2) (Institute for Mar ine Biosciences, Halifax, Canada) An additional source of saxitoxin (STX) utilized was purchased from the National Institute of Standards and Technology Standard Reference Materials (Gaithersberg, MD). All standards were diluted in the same solution th ey were supplied in. Reagents HPLC grade acetonitrile, sodium hydroxide, glacial acetic acid, periodic acid, di sodium hydrogen phosphate, hydrogen peroxide (30% wt. sol), hydrochloric acid (certified ACS), (>99%) formic acid, and ammonium f ormate were all purchased from Thermo Fisher Scientific ( Waltham MA) Glutaraldehyde ( 25 % in water) for algal preservation was also purchased from Thermo Fisher Scientific. All mobile phases for HPLC were filtered through 0.45 m PVDF Millipore filters (Thermo Fisher Sc ientific, Waltham MA) prior to utilization. in house by a Pure Lab Ultra Filtration System (Siemens Water Technologies Corp. Warrendale PA).

PAGE 25

25 Sample Preparation Qualitative Analysis of Filamentous Macroalgae Samples ay within 4 hours of collection and mixed by hand, while rinsing with DI water to remove non algal debris. Once thoroughly mixed, five subsamples (approximately 1 gram wet weight each) were removed from each field collected filamentous macroalgae sample a nd preserved with gluteraldehyde (approximately 0.25%). These specimens were stored at 4C until full algal identi fication/characterization was accomplished. Microscopy was conduc ted by Andrew Chapman at GreenWater Laboratories/CyanoLab in Palatka, FL I n preparation for microscopy, subsamples (approximately one gram wet weight) of preserved samples were placed in plastic petri dishes with sample water or distilled water. A minimum of two wet mounts were made from the subsamples. The entire area of each wet mount was viewed at 100x on a Nikon Eclipse TE200 inverted microscope equipped with phase contrast optics and epi fluorescence. Higher magnifications were used as necessary for species identifications. Identifications were taken to species level whe n possible. A list of observed species was constructed of species organized from most to least abundant based on empirical judgment of the analyst. Epi fluorescence was utilized to aid in characterization of epiphyt es attached to Lyngbya filaments. DNA I solation, Polymerase Chain Reaction, and Sequencing A single filament of Lyngbya wollei was isolated within 4 hours of collection via microscopy from the sample collected in Silver Glen Springs Natural Well (9/29/09). An effort to cut a clean filament abs ent of epiphytic algae was made by observing the

PAGE 26

26 filament at 100x and 400x with phase contrast optics prior to extraction for DNA isolation. The filament was stored at 4 C until DNA isolation could be conducted. DNA analysis was conducted by Mete Yilma z at the University of Florida, Gainesville from a single filament according to Tillet and Neilan (2009 ) The filament was placed in 500 L of XS buffer containing potassium ethyl xanthogenate (1%); Tris HCl, pH 7.4 (100 mM); EDTA, pH 8 (20 mM); ammonium acetate (800 mM); and SDS (1%). The tube was incubated at 70 C for 2 h, vortexed and incubated on ice for 30 min, followed by centrifugation at 15,000 g for 10 min. The supernatant was transferred to a clean microcentrifuge tube and 0.7 volumes of isopro panol was added. DNA was pelleted by centrifugation at 15,000 g for 10 min, washed with 70% ethanol, air dried and resuspended in 50 L of double distilled water. Ampl ification and sequencing of 16sR NA gene was performed as described in Yilmaz et al. ( 2008 ) The p resence of saxitoxin biosynthesis genes was checked by partial amplification of sxtA and sxt G genes with primer pairs designed in this work Primer pair sxt AF ( AGCTGGACTCGGCTTGTTGCTTC) and sxt A R (CACTTGCCAAACTCGCAACAGG) was used to amplify an ap proximately 657 bp fragment within p olyketide synthase sxt A4 region. Primer pair sxt GF ( ATTGAAGCACCAATGGCAGATCG) and sxt GR ( AGAGTTCCGCGTCGGCGAAC) was utilized to amplify an approximate 700 bp fragment within the amidinotransferase gene, sxt G Both PC R reactions contained 5 L of genomic DNA, 20 pmol of each primer (Eurofins, MWG op eron, Huntsville, AL ), 200 M of each deoxynucleoside triphosphate ( Thermo Fisher Scientific, Waltham MA ), 1.5 mM MgCl 2 10 L of 5X green buffer and 2 units of GoTaq Flex i DNA polymerase (Promega Madison WI ) in a total

PAGE 27

27 volume of 50 L Amplification was initiated with denatu ration of the genomic DNA at 95 C for 3 m in, followed by 33 cycles of 95C for 30 sec, 58C for 30 sec, 72 C for 1 min, and end ed with an extension st ep at 72 C for 5 min. PCR products were purified from agarose gels (1.5% w/v) with the QIAquick Gel extraction kit (Qiagen Germantown MD ). Sequencing of 16s r RNA was performed with sequencing primers reported in Yilmaz et al. 2008 and s x t genes were sequ enced on both strands using the same Biotechnology Research core sequencing facility. Sequences were manually checked and corrected using Mega version 4.1 (Tamura e t al. 2007) Toxin Samples Samples were prepared for toxin extraction after qualitative algal and DNA collections were made. Excess w ater was l ightly squeezed from the mats and filaments were cut wet in 1 centimeter ( cm ) lengths to maximize homognization and decrease lyophilizati on time. Samples were frozen ( 20 o C) in freeze flasks and lyophilized at 50 o C (Thermo Savant Modulyo Freeze Dryer System, Thermo Fisher Scientific, Waltham MA) Lyophilized one gram subsets were utilized for extraction experiment s and analysis. Sonication and mechanical homogenization presented difficulties by fouling equipment and /or not sufficiently homogenizing the material. M anually cutting wet filaments prior to lyophilization allowed for the most direct contact of extractio n solvent with the inner cells as determined from microscopy.

PAGE 28

28 P aralytic Shellfish Toxin Extraction Extraction Assessment Methods for extraction of the dried algal material included the AOAC Official Method 2005.06 for PST extraction and by varying the co ncentration and type of acid used. The difficulty when working with algal material such as Lyngbya is effectively extracting the toxin from the thick sheath without extracting accessory pigments that may interfere with analyses and avoiding conversions of the toxins during extraction The two acids assessed were hydrochloric acid (0.1, 0.01, 0.001 M concentrations) and acetic acid (0.1, 0.01, and 0.001 M concentrations). One sample collected fro m Blue Hole (collected 4/2/10 ) was weighed out in 1 gram sub sets and run in triplicate for each extraction. Sample s were extracted in 50 mL of extractant solution with contant stirring for 5 minutes and constant monitoring of temperature. Once cooled to room temperature, the samples were brought ba ck to original volume with the addition of the original extractant solution. All samples were filtered with Whatman Puradisc 0.45 m PVDF syringe filters (Whatman Inc. Piscataway NJ) and stored at 4 o C prior to oxidation, further preparation and analys is. SPE and Lyophilization Assessment A sample set collected in August 2009 was utilized, with Blue Hole, Silver Glen Main stem, Silver Glen Natural Well, and d uplicate collection of Silver Glen Natural Well. Five milliliter aliquots of extractant solut ion s were passed through conditioned 200mg Alltec C18 Solid Phase extraction cartridges ( Grace Analytical, Deerfield IL ) and rinsed with 0.003M HCl (3 mL). The load and wash fractions were collected and frozen in 15 mL glass centrifuge tubes. The fracti ons were lyophilized to dryness and reconstituted in 500 L 0.003 M HCl for analysis with LC/MS at a concentration of 0.2

PAGE 29

29 gram s dried material per mL solution. This solution was diluted to the origina l extract concentration (0.02 g mL 1 ) for comparison wi th original LC/FL and LC/MS Scans Final Extraction Protocol A ll samples from all collections were extracted in the same way with acetic acid One gram dry weight was added to 50 mL of 0. 1 M a cetic acid and the solution boiled for 5 minutes while tempe rature was being monitored closely. Once cooled, the volumes were brought back to original volume with 0.1 M acetic a cid, filtered through Whatman Puradisc 0.45 m PVDF syringe filters and stored at til analysis. Select samples were utilized for pre column oxidation and LC/FL. All samples were analyzed via LC/MS/(MS). Paralytic Shellfish Toxin Analysis HPLC/ Fluorescence Peroxide Oxidation The basic structure of PSTs is a 3,4 propinoperhydropurine tricyclic system ( Table 2 2 ). Oxidation with H 2 O 2 cleaves the propino ring in the presence of an alkaline solution, forming a purine. This purine becomes fluorescent in acidic solution and can be monitored at an excitation of 340 nm and emission of 395 nm ( Figure 2 7 ) (Botana 2008). With separation via HPLC, qualitative and quantitative determinations can be made. The Lawrence, Niedzwiadek Menard method (2005) via prechromatographic oxidation and LC fluorescence is t he accepted AOAC International Official First Action Method due to the successful q uantification of multiple variants of PSTs. This method was used to monitor a wide range of PSTs known to occur, including STX, GTX 5, C1&C2, GTX 2&3, dcGTX2&3, dcNEO & dcSTX.

PAGE 30

30 The peroxide oxidation technique employed was as follows; 25 L 10% (w/v) aqu eous H 2 O 2 was added to 250 L 1 M NaOH and vortexed. One hund red microliters of sample and standards were added to the oxidant solution and allowed to react for 2 min utes at room temperature. Concentrate d glacial a cetic acid (25 L ) was added to acidify t he purine group and allow for fluorescence determination. Standards used to calibrate this method included STX, dcSTX, dcNEO, GTX2&3, C1&2, dcGTX2&3 & GTX5. Oxidation products do eventually degrade, so all oxidized samples were analyzed within 8 hours of their oxidation. Periodate Oxidation Periodate oxidant was prepared fresh daily by mixing 5 mL each of 0.03 M Periodic acid, 0.30 M ammonium formate and 0.3 M Sodium hydrogen phosphate and by adjusting the pH to 8.2 with 175 L 0.2 M Sodium Hydroxide. Al l standards and solutions were oxidized in 100 L aliquots with the addition of 500 L oxidant solution, vortexing and one minute reaction time at room temperature. Five microliters of acetic acid were added to finalize the reaction prior to analysis via LC/FL. Periodate oxidation is utilized to monitor for GTX1&4 & NEO, so standard solutions of GTX1&4 & NEO were prepared via this method. D ue to the significant amount of noise produced by L. wollei extracted samples, non resolved peaks occurring from oxi dation of the PSTs present in the material and the longer run times required to clear the system of retained fluorescent material this method was abandoned for PST characterization LC/MS was utilized to scan for PSTs that would normally be observed in the periodate oxidation procedure (NEO and GTX 1&4).

PAGE 31

31 Chromatographic and Fluorescence Conditions Most PSTs are water soluble non volatile compounds, only existing in ionic forms; therefore HPLC is the preferred method of separation. A Thermo Separations Product (TSP) SpectraSystem High Performance Liquid Chromatography System equipped with a P4000 Pump, AS3000 autosampler controlled by a SN4000 Interface and coupled with a FL3000 for fluorescence detection ( Thermo Fisher Scientific, Waltham MA ) was used for HPLC/FL analysis An Agilent SS420X interface (Agilent Technologies, Inc., Pleasanton CA) was utilized for digitizing analog detector signals from the fluorescence detector and Xcalibur Software ( Thermo Fisher Scientific, Waltham MA ) was utilized for acquisitioning and quantification of PSTs. The LC parameters were modified from the AOAC Official Method 2005.06 to accept a lower flow of 1mL/min and to provide sufficient resolution of PST oxidation products. A 150mm x 4.6 mm, 5 m particle size rever sed phase Aquasil C18 column (Thermo Fisher Scientific, Waltham MA) was used for separation of oxidation products. Two mobile phases where utilized, Mobile Phase A: 0.1 M ammonium formate and Mobile Phase B: 0.1 M ammonium formate in 5% acetonitrile. Bot h were adjusted to pH 6 w/ 6 mL 0.1 Acetic Acid. At a fl ow rate of 1mL/min, runs were 20 minutes with the following linear gradient: 100% 90% A over 5 minutes, hold at 90% A for 5 minutes, from 90% 30% A over 3 minutes, 30% back to 100% over 7 minute ( T able 2 3 ). A 20 L inject ion was used for all LC/FL runs and all samples were monitored both with and without oxidation to verify there were no interfering naturally fluorescent peaks. Fluorescence was monitored with an excitation of 340nm and an emission of 396 nm

PAGE 32

32 Detection Limits Method Detection Limits (MDLs) were determined with a signal to noise ratio of 3:1 in the matrix for all but dcNEO, dcSTX and dcGTX2&3. Standard addition and five point standard curves were used to estimate conservative detec tion limits for dcNEO, d cSTX and dcGTX2&3. The m ethod detection limits are reported in Table 2 4 analyzed with fluorescence with 20 L injections. Standards of dcGTX2&3 were linear to extract concentrations of 15 ppm ( g L 1 ), dcNEO to 5 ppm ( g L 1 ), and dcSTX to 10 ppm ( g L 1 ). Sample Analysis Paralytic shellfish toxins in samples were evaluated by comparing peak areas to standards curves and standard addition via precolumn oxidation and HPLC/F L. A minimum of three point standard curves for dcGTX2&3 an d dcSTX were prepared to bracket sample concentrations, with a s tandard check run at the end of each sample set and verif ied within 10% of initial standard and sample All samples and standards to be analyzed together were oxidized fresh daily with replic ate runs made on standards and on at least one sample per set to assure reproducibility of equipment Fluorescence data was interpreted by normalizing peak areas of samples to standard curves of dcGTX2&3 and dcSTX, which were run daily Resultant values achieved this way were reported in dcGTX3&3 and dcSTX equivalents. HPLC/MS /(MS) Chromatographic and Mass Spectrometry Conditions Liquid chromatography/Mass spectrometry/M ass spectrometry (LC/MS/MS) and LC/MS were utilized for the characterization of PST variants in all samples. A Thermo Finnigan Surveyor HPLC system coupled with a Thermo Finnigan LCQ Advantage

PAGE 33

33 MS n ion trap tandem mass spectrometer was utilized for validation of PSTs in samples. A TSK gel Amide 80 250mm x 2 mm HPLC Column (5 m particl e size) was employ ed for chromatographic separation ( Tosoh Bioscience LLC, Grove City, OH) Two mobile phases were used C: 100% DI with 3.6 mM formic acid and 2 mM ammonium formate and solvent D: 95% (v/v) acetonitrile with 3.6 mM formic acid and 2 mM am monium formate. The elution gradient was as follows; hold at 35% C for two minutes, 35% 70% over ten minutes, 70% held for 5 minutes, return to 35% in 10 minutes and hold for ten additional minutes for re equilibration ( Table 2 5). P aralytic Shellfish Tox in standards utilized in all MS/MS analyses were directly infused on the mass spectrometer and autotuned with Xcalibur software. Once MS/MS parameters and collision energies were optimized, Scans (150 500 m/z), Single Ion Monitoring (SIM) and Selective R eaction Monitoring (SRM) were conducted on standards and samples. Detection Limits Limits of detection could only be calculated for toxins with available standards. Detection limits were determined by standard addition techniques and utilizing a signal to noise ratio of 3:1 with 2 0 L injections for non detectable toxins For toxins positively identified in the matrix, a conservative approach and standard addition was utilized to estimate method detection limits. The elution order of standards was not t he same in the matrix as they were in standard solution due to shifts in retention time as verified via spikes. MDLs and retention times of PSTs for LC/MS and LC/MS/MS are reported in Table 2 6

PAGE 34

34 Sample Analysis Extraction Assessment Samples. All extrac ts from the extraction assessment portion of this study were analyzed with the matrix removed under N 2 gas at 60 C and reconstituted in 0.003M HCl to eliminate/minimize the matrix influence on ionization. Samples were first analyzed for dcGTX2&3 and dcSTX via LC/MS/MS since florescence data exhibited response for only those two toxins. There was a quantitative dis parity between LC/FL data and LC/MS data, so a Total Ion Current (TIC) scan was conducted from 150 500 m/z. Table 2 7 shows ions monitored for with LC/MS /(MS) SCANs and SIM. MS responses from TIC SCAN data was utilized for comparisons between samples fo r LWT 1 6 s ince standards were not available for L. wollei toxins Data utilized for compar ison of dcGTX2 and dcGTX3 was also achieved from SCANs SIM data was utilized for dcSTX comparison. Matrix effects were assess ed with the use of p re and p ost ex traction spikes for STX, dcSTX and dcGTX2&3 Only post extraction spikes were utilized for dcNEO due to the limited availability of standard All collections. All samples extracted in 0.1 M acetic acid were analyzed with the following approach: TIC scans (150 500 m/z ) for dcGTX2, dcGTX3, and L. wollei toxins 1 6 Standard addition with dcGTX2&3 and TIC SCANs were utilized for the quanti tative data achieved for dcGTX2 and dcGTX3 TIC SCAN data resulted in high noise at the dcSTX and dcNEO retention times and could not be used for detection. Instead, SIM was used to monitor for dcSTX (m /z 257) and dcNEO (m/z 273) Standard addition dcSTX with SIM data was utilized for quantification of dcSTX Pre extraction spikes (STX ) and post extraction spikes (STX, dcGTX2&3, dcSTX, dcNEO, GT X1&4 and NEO ) were utilized with representative samples from each

PAGE 35

35 collection site to evaluate retention time shifts and provide additional qualitative data One sample per collection site was also chosen to monitor (SIM) for ion s of toxins with known standards (GTX 1&4 and NEO). To determine reproducibility and handle variability of equipment, analysis was conducted by sample set (sa mple set = one collection month ) and then by sample site A full sample set was analyzed in a si ngle day with replicates to assess intra day variability. Samples analyzed by site were then handled in the same manner with all samples, including field replicates, analyzed in a single day. A sample analyzed at the beginning of a day was re analyzed (SC AN and SIM) at the end of the day. Data for the entire day was accepted i f the rerun analysis was < 1 0% different for all toxins (by comparing MS responses ) Field replicate data was averaged for final concentration calculations and comparison.

PAGE 36

36 Figur e 2 1. Silver Glen Springs Recreational Park photo ; white lines represent ropes restricting general access; Natural Well and Main Stem were sampled in this study. Photo Courtesy of St. Johns River Water Management District

PAGE 37

37 A B C D E F Figure 2 2. S eries of Silver Glen Main Stem p hotos A) Overlooking East Vent a rea B) Red buoys mark the 2 lines separating boat traffic from main spring a rea, where large Lyngbya mat is found (Main Stem) C) Underwater photo of Main Stem mat D) Lyngbya loosely attached to separating rope E) Florida Gar ( Lepisosteus platyrhincus ) in Main Stem F) Lyngbya mat October 2009 G) Lyngbya mat in August 2009 H) Lyngbya mat November 2009

PAGE 38

38 G H Figure 2 2. Continued

PAGE 39

39 A B C D E F Figure 2 3. Series of Silver Glen Na tural Well (Southeastern Vent) p hotos A) Looking into sectioned off area of Natural well from swim area B) Looking out from Natural Well sample area C) Looking down at Natural Well vent D) Standing Lyngbya mat August 2009 E) Lyngbya and Vallisneria americana F) Lyngbya mat January 2010 with Tilapia and nests G) Lyngbya over natural vent August 2009 H) Lyngbya surrounding vent October 2009

PAGE 40

40 G H Figure 2 3. Continued

PAGE 41

41 A B C D E F Figure 2 4. Se ries of Blue Hole (Jug) Spring p hotos A) View of spring from wooden dock B) Cypress Trees with Lyngbya attached loosely to submerged roots C) Looking down B lue Hole Spring Vent D) View of vertically distributed mat loosely attached to tree roots E) & F) View of Lyngbya mat October 2009 G) & H) Lyngbya mat February 2010

PAGE 42

42 G H Figure 2 4. Continued Table 2 1. List of sample collections Site Sampling Dates Silver Glen Main Stem 8/25/ 09* 9/29/ 09 10/23/ 09 11/24/ 09* 1/27/ 10 4/5/ 10 Silver Glen Natural Well 8/25/ 09 9/29/ 09 10/23/ 09* 11/24/ 09 1/27/ 10* 4/5/ 10 Blue Hole Spring 8/27/ 09 9/24/ 09* 10/26/ 09 12/4/09 2/3/ 10 4/2/ 10* D uplicate field replicate col lected at this time

PAGE 43

43 Table 2 2 Saxitoxin structural v ariants Toxin R1 R2 R3 R4 R5 M.W. (Free Base) STX H H H O CONH 2 OH 299.3 GTX1 OH H OSO 3 O CONH 2 OH 411.4 GTX4 H OSO 3 H O CONH 2 OH 411.4 GTX2 H H OSO 3 O CONH 2 OH 3 95.4 GTX3 H OSO 3 H O CONH 2 O H 395.4 GTX5 H H H OCONH SO 3 OH 379.4 NEO OH H H OCONH 2 OH 315.1 C1 H H OSO 3 CONHOSO 3 OH 475.4 C2 H OSO 3 H CONHOSO 3 OH 475.4 dcSTX OH H H OCONH 2 OH 256.3 dcNEO H H OSO 3 CONHOSO 3 OH 272.1 dcGTX2 H H OSO 3 OH OH 352.3 dcGTX3 H OSO 3 H OH OH 35 2.3 LWT 1 H H OSO 3 OCOCH 3 H 37 8.3 LWT 2 H H OSO 3 OCOCH 3 OH 39 4.3 LWT 3 H OSO 3 H OCOCH 3 OH 39 4.3 LWT 4 H H H H H 24 0.3 LWT 5 H H H OCOCH 3 OH 29 8.3 LWT 6 H H H OCOCH 3 H 28 2.3 *Structures acquired from Onodera et al. 1997, there are no know n standards for these variants

PAGE 44

44 Figure 2 5. Conversion of saxitoxin to fluorescent d erivative for HPLC/FL detection

PAGE 45

45 Table 2 3 LC/FL g radient Time (min) Solvent A Solvent B 0 100% 0% 5 90% 10% 10 90% 10% 13 30% 70% 20 100% 0% Table 2 4 MD Ls for PSTs via f luorescence in elution order ( g g 1 dry weight ) Toxin MDL dcGTX 2&3 0.5 C1&C2 0.5 dcSTX 0.5 dcNEO 5.0 GTX 2&3 2.5 GTX 5 5.0 STX 1.0 Table 2 5 LC/MS g radient Time (min) Solvent C Solvent D 0 35% 65% 2 35% 65% 12 70% 30% 17 70% 30% 27 35% 60% 37 35% 65%

PAGE 46

46 Table 2 6 MDLs and r etention times of PSTs analyzed via LC/MS/(MS ) Toxin SIM MDL ( g g 1 ) SRM MDL ( g g 1 ) Retention time (min) Matrix Retention Time (min) Standard STX 5 10 13.5 16.5 dcSTX 5 13 13.0 16.7 dcNEO 20 25 13.1 18.8 dcGTX2 19 100 8.0 7.6 dcGTX3 38 100 8.3 7.9 GTX1 10 8.1 3.6 GTX4 12 8.5 3.6 NEO 9 16.9 13.1 LWT 1 6.2 LWT 2&3 6.6 LWT 4 3.0 LWT 5 12.0 LWT 6 11.8

PAGE 47

47 Table 2 7 Molecular and fragmentation i ons monitored LC/MS Toxin [M+H] + SIM ion Product Ions STX 300 300 282 266 221 204 dcSTX 257 257 239 222 dcNEO 273 273 255 225 dcGTX2 353 273 273 255 dcGTX3 353 353 335 273 255 GTX1 412 332 GTX4 412 412 NEO 316 316 *LWT 1 379 379 *LWT 2 395 395 *LWT 3 395 395 *LWT 4 241 241 *LWT 5 29 9 299 *LWT 6 283 283 N o known standards available, data from Onodera et al. 1997 Note: All other data adopted from Dell'Aversano et al. 2005 except product ions (determined from autotune, Xcalibur software)

PAGE 48

48 CHAPTER 3 RESULTS Field Data Water Quality Water quality data collected during this study (August 2009 April 2010) is outlined in Table 3 1 Temperatures remained constant throughout the study p eriod, ranging no more than 0.2 ported average values at each site Blue Hole Spring averaged 21.37 while Silver Glen Springs averaged 23.27 In the Silver Glen Main Stem, pH ranged from 7.85 8.02 while in the Natural Well Vent pH was slightly lower ranging from 7.72 7.86 Blue H ole Spring pH was 7.48 7.59. Silver Glen Springs is naturally influenced by saline ground water inputs so salinity and Specific Conductivity ( S.C. ) measurements are higher than that of Blue Hole. Salinity was highest from the Natural Well Vent (1.016 p su ), just slightly lower in the Main Stem (0.938 p su ) and much lower in Blue Hole (0.142 p su ). Specific conductanc e measured at the Natural Well v ent of Silver Glen Springs was the highest at 1982 S cm 1 slightly lower at the main stem (1858 S cm 1 ) and low at Blue Hole (298 S cm 1 ). The data collected at the time of study are comparable to data from Florida Department of Environmental Protection ( FDEP ) Dissolved oxygen levels are low in all th e spring systems near vents. T hey were higher during this study at Silver Glen Springs both Main S t em (3.5 4.06 mg L 1 ) and Natural W ell areas (3.1 3.5 mg L 1 ) than that of Blue Hole Spring (1.8 2.2 mg L 1 ). This represents an increase from DO levels reported by the FDEP from 2001 2006 where the average in Sil ver Glen Springs was 2.78 mg L 1 and 1.83 in Blue Hole Spring The overstory densities were quite different from site to site, potentially impacting the

PAGE 49

49 Lyngbya mat growth an d health. In Silver Glen, the Main S tem collections were free from canopy covera ge, representing an overstory density of 0%; trees lining the embankment did not have any influence over the large mat in the middle of the spring. Conversely, the mat sampled at Natural Well was surrounded by an embankment with overstory density varying from 7 21%, and an average of 14%. Blue Hole represented the highe st canopy coverage with a range from 69 89%, an average of 81% coverage. Samples collected at the top of the mats in Silver Gle n Springs averag ed 1.1 m eters in depth at the Main S tem and 0 .97 meters in the Natural Well. The collection s from Blue Hole were obtained at the max depth of 1 m eter so all collections for all spring samples were roughly at the same depth of 1 meter The Florida Department of Environmental Protection (FDEP) prod uced a n Initiative R eport in 2010 summarizing water quality data for over 49 spring vents (Harrington et al. 2010) The data released in that report is utilized here. Water quality for the Main Vent in Silver Glen Springs was sampled 21 times over a per i od from 2001 2006. Blue Hole S pring was sampled 18 times from 2001 2002 and 2004 2006. Data collected for the F DEP study (2001 2006) are reported in Table 3 2 The FDEP Initiative Report states that Silver Glen Springs has nitrate levels near or at the r ange that would be considered true background levels, with Nitrate + Nitrite lev els at 0.05 mg L 1 Conversely, nitrate levels in Blue H ole are averaged to be 0.695 mg L 1 which is higher than the FDEP proposed surface water standard for nitr ogen in spri ng vents of 0.35 mg L 1 Sulfate, calcium, potassium, and Total Dissolved Solid levels are higher in Silver Glen Springs than that of Blue Hole Spring but turbidity, Color, and pH values are similar. Fecal coliforms were not detected at either spring.

PAGE 50

50 Field Observations Silver Glen Springs Lyngbya wollei appeared to have the most observed changes during this study with regard to the mat size and color. The average thickness of the mat (from top of the mat to max depth) was 12 cm (0.12m) in August 2009 ( Table 3 3) The mat remained between 12 cm and 19 cm until January 2010 where it gained over 10 cm for average thickness to 30 c m. One last sample period in April 2010 resulted in a m easured average mat thickness of 46 cm. Other observations included t he color of the mat, which seemed to change from a burnt brown/orange color in the summer to a darker black color in the wint er months. Til apia created deep beds within the mats for breeding purposes in January 2010 and Atlantic S tingray were abundant dur ing the summer months. Mullet were abundant year round and were frequently observed picking up the Lyngbya filaments and either ingesting it or just the epiphytes covering the Lyngbya It is unknown at this time if mullet actually fully ingest Lyngbya a study examining the stomach contents of the fish would be required to determine that. Silver Glen Natural Well Lyngbya appeared the darkest in color of the three sites The mat was thickest in November of 2009, an average of 32 c entimeters (c m ) although it appeared to decline in growth when sampling was conducted in April 2010, with an average thickness of 16 cm ( Table 3 3) The thickness of the Lyngbya mat at Blue Hole Spring was not measured since it was vertically distributed and the mat was intertwi ned with the roots of a cypress tree although Lyngbya was present along each point of the transect. Algal Analysis A ll 24 samples collected were dominated by the filamentous cyano bacterium Lyngbya wollei Farlow ex Gomont (syn. Plectonema wollei ) as determ ined via

PAGE 51

51 microscopy Lyngbya wollei has been disputed in the literature for some time now, as molecular and morphological evidence is sometimes confounding (Joyner et al. 2008). The mats collected here were determined to be L. wollei by meeting criteria published in Speziale and Dyck (1992). In some samples collected, other Lyngbya species were identified ( Lyngbya cf. aestuarii & Lyngbya cf. major ), but only made up a small fraction of the entire sample. All collections were comprised of more than one a lga, but the dominant algal species in every sample was L. wollei (Figure 3 1) Silver Glen Main Stem yielded t he highest algal species list, followed by Blue Hole Spring and then Silver Glen Natural Well. Table 3 4 corresponds to qualitative data acquire d from all samples. F ine oscillatorialean filaments reported in Table 3 4 were primarily Heteroleibleinia sp. with some Leibleinia sp. (epiphytic) and Leptolyngbya sp. (metaphytic). The primary epiphytes identified were the previously mentioned oscillato rialean filaments, the pennate diatom Cocconeis sp. and the chroococcalean cyanophyte Chamaesiphon sp. Micrographs of commonly found epiphytes are in Figure 3 2. Changes in the amount and types of epiphytes on the L. wollei did occur from month to month, but were not quantitated. Molecular Biology A 1366 bp near full le ngth16S rR NA gene fragment was recovered. A blast search with other Lyngbya sequences in the National Center for Biotechnology Information ( http ://www.ncbi.nlm.nih.gov/ ) database revealed highest identities to two L. wollei 16S rR NA sequences. While the sequence was 95% identical to L. wollei strain Carmichael/Alabama (Accession: EU439567) over 1353 nucleotides, it was 96% identical to another L wollei sequence (Accession: EU603708) over 726 nucleotides, both of which are saxitoxin producers (Kellman et al. 2008, Mihali et al. 2009) Identities

PAGE 52

52 t o other Lyngbya sequences in the database ranged from 86% to 92%. A blast search including other species showed a Blennothrix sp. strain having 98% identity to our sequence over 1107 bp (Accession: EU586735). However the identity of 16S rR NA sequence of this Blennothrix strain to an other Blennothrix sequence (Accession: EU253968) in the NCBI database was around 89%. Although the 16S rR NA sequence of this filament shows a high identity (98%) to a sequence presumably obtained from Blennothrix we suggest t hat this is pro bably a misidentified species. Moreover we could not find a published article on this sequence and there are no available morphological descriptions for this species from which the sequence was obtained. The partial sxt A gene sequence obt ained from the single L. wollei filament showed 97% identity to the corresponding region in L. wollei strain Carmichael/Alabama (Accession: EU629174), while identities to other sxt A genes was lower from 90% to 92% from saxitoxin producing species such as C raciborskii, Aphanizomenon sp ., Aphanizomenon flos aquae, Aphanizomenon gracile, Aphanizomenon issatschenkoi, Anabaena circinalis, Anabaena planktonica and Anabaenopsis elenkinii A 672 bp sxt G sequence was recovered from the L. wollei filament which sh owed 98% identity to that of L. wollei strain Carmichael/Alabama (Accession EU629180). Identities to other sxt G sequences from other species were lower from 93% to 96%. HPLC/Fluorescence Peroxide Oxidation Products Precolumn peroxide oxidation provided go od sensitivity to PST standards and reproducible oxidation products for f luorescence. Sample extractions of samples from Blue Hole Spring with varying concentrations of acid were initially characterized with this

PAGE 53

53 method. Figure 3 3 represents a typical c hromatogram acquired from pre column peroxide oxidation of a mixed standard solution. Extraction Assessment Initial assessments of Lyngbya extracts were made with regard to differences in Colored Dissolved Organic Matter (CDOM) and pH. S ince pigments ca n interfere with analyses it was important to utilize an extraction protocol that p rovided a representative toxin profile but d id not result in high concentrations of pigment. Extractions with 0.1 M hydrochloric acid resulted in the highest concentration of extracted pigments a very dark purple extract followed by 0.001 M HCl All other extractions, 0.01 M HCl and all acetic acid extractions, resulted in clear solutions with very little pigment co extracted. This was also observed in flu orescence chrom atograms, where all non oxidized samples exhibited multiple peaks but there was a noticeable increase in peak size at a retention time of 16.8 minutes in n on oxidized samples extracted with 0.1 M HCl and 0.001 M HCl. Although these peaks do not elute at retention times that would interfere with known toxin standards, samples with higher concentrations of pigment present ed problems when attempting to concentrate the extract ( lyophilization SPE ) pH, an important factor with regard to conversions and degr adation of PSTs, was measured both before and after boiling the samples in the extraction solution to verify they remained acidic ( Table 3 5 ) Samples extracted with varying concentrations of acids were assessed using LC/FL and later by LC/MS. Only two major peaks were observed in LC/FL chromatograms of peroxide oxidized samples in all extractions retention times matched those of dcGTX2&3 and dcSTX /dcNEO Peaks corresponding to standards of STX, C1&2, GTX 2&3 and GT X 5 were not obse rved. Figure 3 4 re presents common

PAGE 54

54 chromatograms achieved through peroxide oxidation LC/FL experiments conducted on Lyngbya samples. All extractions exhibited accep t able reproducibility with respect to replicates (<10% difference) and spike responses. Variations in acid t ype and concentration result ed in changes to peak size and FL detector response. Peak areas of the triplicate extractions and replicate runs were averaged and compared to standard curves of dcGTX23 and dcSTX. Figure 3 5 summarizes the differences in extr actions. As acid concentration went down, so did the peaks corresponding to both dcGXT2&3 and dcSTX. This could be due to many reasons, including inefficiency of sample extraction as the acid becomes more dilute, degradation of toxins in higher pH (less likely) and changes in the profile which provide a different response once oxidized and analyzed. Even though the resultant peaks may have originated from multiple toxins (e.g. L. wollei PST toxins), it is theorized that the oxidation products are the sa me. However, q uantification could not be achieved utilizing this method since stan d ards of L.wollei toxins are not available their optimal oxidation conditions are unknown and the resultant response via fluorescence may not be equal. In addition, some known PST toxins also produce similar, if not the same oxidation products, and cannot be distinguished with this method For instance, separation of dcNEO and dcSTX is not possible with precolumn oxidation and peak detection at retention times correspondi ng to oxidation products of either toxin cannot be fully characterized or quantitated since they exhibit a different fluorescent response at the same concentrations Other t oxins known to react this way include dcGTX2 and dcGTX3 which are typically combi ned in

PAGE 55

55 pre column oxidation since the oxidation products cannot be fully resolved with this method It should be noted that s amples were re analyzed two weeks after being stored at 4 C, and although responses to samples extracted with acetic acid and mo re dilute hydrochloric acid remained unchanged; the 0.1 M HCl acid extraction represented a 40 % increase in the dcGTX2 & 3 eq and a 16 % increase in dcSTX eq peaks This could be due to the continued conversion s or changes in toxin profile due to hydrolysis as observed in other PST matrices in other studies. Pre column peroxide oxidation does not allow for specificity with regard to many toxins, including L. wollei toxins. This is an inherent problem with pre column oxidation of PSTs and would require furthe r work to improve upon the approach. Further characterization of extraction products was conducted by LC/MS. SPE and Lyophilization Assessment Solid Phase Extraction with C18 resulted in lowered fluorescence response Th ere was an average recovery of 8 8% (81% 92%) for the dcGTX2&3 peak and 83% (71% 90%) for the dcSTX peak. The spiked samples of STX resulted in a n average recovery of 117% when compared to standard curves Initial lyophilization efforts resulted in significant losses, up to 80% for som e samples ( LC/FL ). The removal of pigments and other compounds co extracted was necessary to increase recoveries and allow for reconstitution of samples in less volume Post C18 extracted material was lyophilized and final recoveries via LC/FL were avera ged 67% (62% 75%) for the dcGTX2&3 peak and an increase in response was observed for the dcSTX peak, with an average of 104% (98% 107%) from the original extract. The saxitoxin spike recovery was 75% post SPE and lyophilization.

PAGE 56

56 Sample Analysis Samples analyzed via pre chromatographic peroxide oxidation LC/FL included collections from all spring sites in the month of August 2009 and April 2010. Other samples were not analyzed with this method since qualitative data from LC/FL is limited and chromatograp hic separation of L. wollei toxins could not be achieved. As a screening method for toxins with good chromatographic separation (i.e. STX, C1&C2, GTX2&3, GTX 5), it provided a useful tool for add itional confirmation It also provided supporting extractio n efficiency data when used wi th pre extraction STX spikes, as exhibited in Figure 3 6 HPLC/MS Extraction Assessment Samples extracted with varying acids analyzed by LC/MS exhibited similar re sponses with regard to L. wollei toxins and decarbamoyl toxins with the exception to the 0.1 M HCl extracted samples. Figure 3 7 summarizes the MS response s to toxins detected in the varying acid extractions. Data from both LC/F L and LC/MS extractions supported the relationship between acid concentrations and tota l tox ins extracted, with exception to the 0.1 M Hydrochloric acid extractions. The stronger HCl extraction resulted in only the detection of dcGTX2, dcGTX3, and dcSTX. The L. wollei toxins were not detected. This is potentially due to the conversion of the L. wollei toxins to dcGTX2&3 and dcSTX via hydrolysis. This was further supported by an increase in dcGTX2 response (812 1122%, average 977%), dcGTX3 response (459 34 73%, average=1654%) and dcSTX MS/MS response (154 2689%, average 929%) in the 0.1 M H Cl extracted material versus all other acid extractions ( examples are in Figure 3 8 and 3 9 )

PAGE 57

57 The use of more dilute acids resulted in the detection of all the L. wollei toxin parent ions, dcGTX2&3 and dcSTX. Due to potential conversions, hydrochloric a cid was avoided for final extracti ons on all samples and a 0.1 M a cetic acid solution was chosen instead The acetic acid extractions resulted in accep t able recoveries of pre extraction and post extraction spikes, exhibited good reproducibility between ex tractions (<10% difference) and allowed for the detection of all L.wollei toxins, dcSTX and dcGTX2&3. SPE and Lyophilization Assessment The use of SPE alone in an attempt to clean the sample matrix did not provide an increase in signal to noise and did n ot prevent shifts in retention time The use of SPE in conjunction with lyophilization for samples analyzed via LC/MS resulted in significant levels of noise for both scans and MS/MS. Although LC/FL data showed that there was >62% recovery of oxidized PS T products, lyophilization of the extracts post SPE to achieve a higher extract concentration resulted in complete ion suppression and toxins ( L.wollei toxins, STX, dcSTX, dcGTX2&3) were not detectable Sample Analysis All samples extracted with 0.1 M Acet ic acid and analyzed with LC/MS (SCANs and SIM) contained decarbamoylgonyautoxin(s), decarbamoylsaxitoxin, and LWTs 1 6. All other toxins were not present or below the detection limits established in this study, including dcNEO, GTX1, GTX4 and NEO. Some samples did indicate a presence of dcNEO via SIM but the levels were below the MDL and could not be verified Analysis with SIM indicated the presence of GTX1 as well, but LC/MS/MS was not conducted at this time to confirm the presence. Further method d evelopment and analysis must be conducted to establish the presence/absence of GTX1 and dcNEO.

PAGE 58

58 Matrix effects were apparent for all standard spikes and included retention time shifts and reduced MS response to standards Retention times for dcGTX2&3 shi fted the least (<1 min) but dcSTX dcNEO, NEO, GTX1&4 peaks were detected up to 5.7 minutes prior to standard retention times without the matrix. LC/MS/MS was utilized to verify retention times of standard spikes. Field replicate data represent ed signif icant variability for some toxins and samples. D ifferences greater than 20% were observed f or some samples and even a 61 % difference for LWT 6 collected 1/27/10 from SGS N atural Well. The highest variability in all duplicate analyses was observed with LW T1 and LWT 6 data (both non toxic LWTs) This level of variability was not observed in replicate runs of the same sample (<10% difference) or in triplicate extractions of the same sample so it may be concluded that variability was inherent in the collecti ons themselves and the L. wollei mat was not homogeneous in nature F ield replicate data was averaged to repres ent data from any given site. Decarbamoylgonyautoxin 2&3 concentrations for all sites and sampling months are repor ted in Table 3 6 The Main Stem of Silver Glen Springs represented the lowest concentrations of dcGTX2&3 while SGS Natural Well had the highest reported concentrations for dcGTX2 and dcGTX3. However, Blue Hole had the highest average concentrations for both toxins with the lowest v ariability between sample dates (standard deviation= 7 ; n=6 ). Concentrations of dcGTX2 ranged from not detectable (<5 g g 1 ) to 43 g g 1 in Silver Glen Main Stem and dcGTX 3 concentration range d from 10 g g 1 to 19 g g 1 Silver Glen Natural Well dcG TX2 concentrations ra nged from 15 to 62 g g 1 and dcGTX3 ranged from 12 to 29 g g 1 Blue Hole Spring

PAGE 59

59 concentrations of dcGTX2 ranged from 31 to 54 g g 1 and dcGTX3 ranged from 16 to 29 g g 1 The lowest concentrations of dcGTX2&3 for all sites were in the month of Se ptember. The sum of both dcG TX2 and dcGTX3 are shown in Figure 3 10 over the entire sampling period. Decarbamoyl saxitoxin (dcSTX) concentrations for all sites ranged from 16 33 g g 1 with the highest concentration reported in the mont h of January collected at SGS Main Stem. The lowest concentrations of dcSTX for both Silver Glen samples were in the month of September while it was lowest in February for Blue Hole Conversely, the highest concentrations for Silver Glen were in January collections. Overall the concentration of dcSTX did not vary greatly for any sampling site over the entire collection period with an average of 24 g g 1 (stdev= 4.9 %CV= 20 % ). Values for dcSTX concent ration are reported in Table 3 7 A summary of re sul ts can be seen in Figure 3 11 Lyngbya wollei toxins were compared by MS response s only since standards were not available for quantification Elution order of LWT toxins was not the same as in previous studies, instead, the order of elution in this stud y of L. wollei toxins was: LWT 4, LWT 1, LWT 2&3, LWT 6 followed by LWT 5. Decarbamoyl toxins dcGTX3 and dcGTX2 eluted after LWT 2&3 while dcSTX eluted between dcGTX2&3 and LWT 6. No other toxins were detected above detection limits in samples. Lyngbya wollei toxin s 2 and 3 are isomers and have the same molecular weight of 396 ([M+H + ] = 395). Resolution of the two toxins was not apparent in this study, so they were combined as LWT 2&3. SCANs did exhibit a peak at the same RT of LWT 2&3 with m/z 315,

PAGE 60

60 re presenting the loss of a sulfate group and providing additional confirmation of LWT 2/3 presence. Samples spiked pre extraction with saxitoxin (10 g) were analyzed via MS/MS and it was discovered that LWT 5 fragments in the same way as saxitoxin even tho ugh its parent m/z is 299 (saxitoxin m/z is 300). In addition, the matrix shifted saxitoxin spikes to the same retention time as LWT 5, not allowing for resolution or the ability to differentiate between the two. Even when MS parameters were set to the s trictest possible settings, interference was still apparent, so STX extracted spikes were not utilized as a measure of extraction efficiency for LC/MS However the fragmentation pattern observed for LWT 5 provided additional confirmation of its presence and structure as well as an MS/MS program for analysis of the toxin in L. wollei samples. Representative SCAN data showing SIM ions for L. wollei toxins from each si te can be viewed in Figure s 3 12 to 3 14 L. wollei toxin 4, a non toxic form of PST resu lted in the highest responses for all sites and sampling dates Data was average d from each sampling mont h and represented in Figure 3 15 to show that the toxin profile was similar for each site. Figure s 3 16 to 3 20 represent MS response for all Lyngbya wollei toxins detected at each site from month to month as determined form TIC chromatograms The data is derived from MS responses only since standards for quantification were not available. This limits the ability to make comparisons between toxins or infer which PSTs may dominate the profile. Although it appears that LWT 4 dominated the profile in all extractions, the response was not calibrated and actual toxin levels could not be determined.

PAGE 61

61 Table 3 1. Water quality and sample collection d ata f or sample sites (August 2009 April 2010) (a veraged data, n=6) Parameter Blue Hole Silver Glen Main Stem Silver Glen Natural Well Average S.D. Average S.D. Average S.D. T emperature 21. 6 0.1 23. 2 0.1 23.0 0.0 pH (s.u.) 7.5 0.0 7.9 0.1 7.8 0.1 S alinit y (psu) 0.14 0 .00 0.9 4 0.02 1.02 0 .00 Specific Conductance (S cm 1 ) 298 9 1858 26 1982 22 Dissolved Oxygen (mg L 1 ) 2. 1 0. 2 3. 8 0.2 3.2 0.2 Average Canopy Coverage % 81% 7% 0% 0 % 14% 5% Average Collection Depth (m) 1.0 0 0.00 1. 07 0.20 1.00 0.19 Note: S.D. = Standard Deviation

PAGE 62

62 Table 3 2 Water quality d ata for Blue Hole Spring and Silver Glen Springs from FDEP Initiative Report 2011, collections from Silver Glen Springs only made at one site, the main pool area Blue Hole Spring Silver Glen Springs Parameter Average Range Average Range Nitrite & Nitrate (mg L 1 ) 0.695 0.510 0.790 0.0 49 0.046 0.055 Ammonia (mg L 1 ) 0.005 0.005 0.005 0.005 0.005 0.005 Orthophosphate (mg L 1 ) 0.047 0.039 0.047 0.027 0.024 0.028 Potassium (mg L 1 ) 0.3 6 0.34 0.41 8.7 0 7.20 9.40 Sodium (mg L 1 ) 3 .0 2.7 3.0 252 .0 216 .0 257 .0 Chloride (mg L 1 ) 4.9 4.3 5.5 450 380 480 Specific Conductance (S cm 1 ) 298 287 309 1905 1540 2050 Sulfate (mg L 1 ) 4.7 4.2 28.0 170 150 180 Dissolved Oxygen (mg L 1 ) 1.83 1.26 2.76 2.78 2.03 4.10 Calcium (mg L 1 ) 5 4.0 47.9 56.8 73.0 62.7 79.1 Total Dissolved Solids (mg L 1 ) 168 150 164 1020 854 1060 Turbidity (mg L 1 ) 0.15 0.03 1.40 0.10 0.03 0.30 TOC (mg L 1 ) 0.5 0.5 1.3 0.5 0.5 1.3 Color (Pt co) 2.5 2.5 5.0 2.5 2.5 5.0 pH (su) 7. 50 7.1 4 7.72 7.74 7.42 7.92 Fecal Coliforms (col 100mL 1 ) 0 0 0 0

PAGE 63

63 Table 3 3 Mat thickness measured at Silver Glen Springs 5 collections per sample date SITE DATE Mat Thickness (cm) Average Silver Glen Springs Main Stem 8/25/ 09 20 10 10 10 10 12 8/25/09 20 10 10 10 10 12 9/29/09 15 20 20 20 20 19 10/23/09 10 10 15 20 20 15 11/24/09 30 20 10 10 10 16 11/24/09* 30 35 10 10 10 19 1/27/10 40 10 40 40 20 30 4/5/10 40 50 50 40 50 46 AVERAGE 21 Silver Glen Springs Natural Well 8/25/ 09 10 10 5 25 5 20 9/29/09 20 10 5 5 40 16 10/23/09 10 30 10 40 10 20 10/23/09 10 10 10 10 10 10 11/24/09 20 40 30 40 30 32 1/27/09* 30 30 10 10 10 18 1/27/10 35 35 30 20 10 26 4/5/10 5 30 30 10 5 16 AVERAGE 20 *Denotes field duplicate colle ction

PAGE 64

64 Table 3 4 Qualitative a lga e l ist from all samples DATE Silver Glen Main Stem Silver Glen Natural Well Blue Hole Spring August 2009 Lyngbya wollei Lyngbya wollei Lyngbya wollei fine oscillatorialean filaments spp. fine oscillatorialean fila ments spp. fine oscillatorialean filaments spp. Spirogyra sp. 1 Phormidium sp. 2 Microchaete sp. 1 Phormidium sp. 3 Oedogonium sp. 1 Stigeoclonium sp. 1 Oedogonium sp. 2 Ulva flexuosa Cladophora glomerata Anabaena sp 1 oscillatorialean filament sp. 1 Pseudanabaena sp. 1 September 2009 Lyngbya wollei Lyngbya wollei Lyngbya wollei fine oscillatorialean filaments spp. fine oscillatorialean filaments spp. fine oscillatorialean filaments spp. Stigeoc lonium sp. 1 Spirogyra sp. 1 Microchaete sp. 1 Phormidium sp. 3 Oedogonium sp. 1 Phormidium sp. 2 Anabaena sp. 1 Batrachospermum sp. 1 Microchaete sp. 2 Mougeotia sp. 1 October 2009 Lyngbya wollei Lyngbya wollei Lyngbya wollei fine os cillatorialean filaments spp. fine oscillatorialean filaments spp. Phormidium sp. 3 Anabaena sp. 1 Homoeothrix sp. fine oscillatorialean filaments spp. Spirogyra sp. 1 Phormidium sp. 2 Oedogonium sp. 2 Microchaete sp. 1 Stigeoclonium sp 1 Oedogonium sp. 1

PAGE 65

65 Table 3 4 Continued DATE Silver Glen Main Stem Silver Glen Natural Well Blue Hole Spring Nov./Dec. 2009 Lyngbya wollei Lyngbya wollei Lyngbya wollei fine oscillatorialean filaments spp. fine oscillatorialean filament s spp. Phormidium sp. 2 Lyngbya cf. aestuarii fine oscillatorialean filaments spp. Lyngbya cf. major Homoeothrix sp. 1 Phormidium sp. 1 Phormidium sp. 2 Anabaena sp. 1 Phormidium sp. 4 Phormidium sp. 2 Microchaete sp. 1 Rhizo clonium hieroglyphicum Stigeoclonium sp. 1 Oedogonium sp. 1 Spirogyra sp. 1 Jan./Feb. 2010 Lyngbya wollei Lyngbya wollei Lyngbya wollei fine oscillatorialean filaments spp. fine oscillatorialean filaments spp. fine oscillatoriale an filaments spp. Lyngbya cf. aestuarii Phormidium sp. 2 oscillatorialean filament sp. 2 chlorophyte filament sp. 1 Microchaete sp. 1 Mougeotia sp. 1 April 2010 Lyngbya wollei Lyngbya wollei Lyngbya wollei fine oscillatorialean fila ments spp. fine oscillatorialean filaments spp. fine oscillatorialean filaments spp. Stigeoclonium sp. 1 Phormidium sp. 3 L. aestuarii (Mertens) Liebmann sensu Prescott 1962; unrevised in Komarek and Anagnostidis

PAGE 66

66 A B C D E Figure 3 1. Series of m icrographs of Lyngbya wollei A) Exhibiting false branching from Blue Hole Spring with phase contrast optics 400x B) False branching from Silver Glen Springs Natural Well 400x C) Silver Glen Main Stem fi lament, b right field 400x D) Filament fro m Blue Hole E) Multiple filaments at 100x from Silver Glen Natural Well

PAGE 67

67 A B C D Figure 3 2. Series of micrographs of Lyngbya wollei epiphytes, all at 400x A) Leibleinia sp. on Lyngby a viewed with epifluorescence from Silver Glen Springs Main B) Epifluorescent image of empty L. wollei sheath with attached cyanobacterial epiphytes collected Silver Glen Main C) Heteroleiblenia sp. attached to Lyngbya from Silver Glen Natural Well D) Co cconeis sp. attached to Lyngbya filament from Silver Glen Natural Well E) Chamaesiphon sp. attached to filam ent from Silver Glen Main Stem F) Cyanophyte epiphytes from Blue Hole G) Heteroleibleinia sp. from Silver Glen Main Stem H) Cocconeis on Lyngbya from Blue Hole

PAGE 68

68 E F G H Figure 3 2 continued

PAGE 69

69 Figure 3 3. Fluorescence chromatogram of standard solution mixture of dcGTX &3, dcSTX, GTX 2&3, GTX 5, and STX (dcNEO not shown as it co elutes with dcSTX) Table 3 5 pH of each extraction solution before and after boiling Extraction pH before pH after 0.1 M HCl <1.5 4 .0 0.01 M HCl 2.5 4.5 0.001 M HCl 5 .0 6 .0 0.1 M Acetic Acid 3.5 3.5 0.01 M Acetic Acid 4.5 5.5 0.001 M Acetic Acid 5 .0 6 .0

PAGE 70

70 Figure 3 4 Fluorescence chromatogra m of oxidized extraction (0.1 M acetic a cid) with peaks appearing at retention times the same as dcGTX2&3 and dcSTX/dcNEO (Silver Glen Main Stem collected 8/25/09)

PAGE 71

71 Figure 3 5 Extraction technique d ata (HPLC/FL) Data is repre sented by normalizing peaks to standards of dcGTX2&3 and dcSTX

PAGE 72

72 Figure 3 6 Fluorescence chromatogra m of oxidized extraction (0.1 M acetic a cid) with peaks appearing at retention times the same as dcGTX2 &3 and dcSTX and with 10g STX s piked pre extract ion (Blue Hole collected 8/27/09)

PAGE 73

73 Figure 3 7 Extraction t echnique d ata (HPLC/MS). Includes L. wollei toxins, dcGTX2 dcGTX3 and dcSTX relative MS responses

PAGE 74

74 A B Figure 3 8 HPLC/MS chromatograms of dcGTX2 & dcGTX3 SIM ions from Blue Hole Spring extracted with two differen t acids A) 0.1 M hydrochloric acid. B) 0.1 M acetic a cid

PAGE 75

75 Figure 3 9 HPLC/MS/MS chromatograms of dcSTX from Blue Hole Spring extracted with two different acids (0.1 M hydrochloric a cid and 0.1 M acetic a cid)

PAGE 76

76 Table 3 6 Decarbamoylgonyautoxin 2&3 (dcGTX2&3) concentrations ( g g 1 ) for all sites and sampling months Month Collected SGS Main Stem SGS Natural Well Blue Hole dcGTX2 dcGTX3 dcGTX2 dcGTX3 dcGTX2 dcGTX3 August ND 11 40 26 30 25 September ND 9 ND 17 29 23 October ND 16 44 26 38 16 Nov/Dec ND 17 60 28 49 21 Jan/Feb 43 17 26 17 51 28 April 21 8 22 11 41 24 Note: ND = not detectable above detection limit Figure 3 10 Sum of dcGTX2 and dcGTX 3 concentrations for all sites over sampli ng period ( reported in g g 1 dry weight of lyophilized L. wollei ) Error bars represent standard error.

PAGE 77

77 Table 3 7 Decarbamoylsaxitoxin (dcSTX) concentrations ( g g 1 ) for all sites and sampling months Month Collected SGS Main SGS Natural Well Blue Hol e August 22 25 25 September 17 16 28 October 22 24 30 Nov/Dec 31 28 25 Jan/Feb 33 29 17 April 25 23 19 Figure 3 11 Concentrations of dcSTX for all sites over sampling period ( g g 1 dry weight of lyophilized L. wollei ) Error bars represent st andard error

PAGE 78

78 Figure 3 12 HPLC/MS SCAN of Silver Glen Springs Main Stem (collected 1/27/10) with SIM ions for L. wollei Toxins 1 6

PAGE 79

79 Figure 3 13 HPLC/MS SCAN of Silver Glen Springs Natural Well (collected 11/24/10) with SIM ions for L. wollei Toxin s 1 6

PAGE 80

80 Figure 3 14 HPLC/MS SCAN of Blue Hole (collected 4/2/10) with SIM ions for L. wollei Toxins 1 6

PAGE 81

81 Figure 3 15 Averaged MS responses over period of study for L. wollei T oxins 1 6 at each site ( error bars represent sample standard deviation )

PAGE 82

82 Figure 3 16 MS Responses to Lyngbya wollei T oxin # 1 detected from all sites o ver the sampling period Error bars represent sample standard deviation.

PAGE 83

83 Figure 3 17. MS Responses to Lyngbya wollei Toxin #2&3 detected from all sites o ver the sampling per iod. Error bars represent sample standard deviation.

PAGE 84

84 Figure 3 18. MS Responses to Lyngbya wollei Toxin #4 detected from all sites o ver the sampling period. Error bars represent sample standard deviation.

PAGE 85

85 Figure 3 19. MS Responses to Lyngbya wollei Toxin #5 detected from all sites o ver the sampling period. Error bars represent sample standard deviation.

PAGE 86

86 Figure 3 20. MS Responses to Lyngbya wollei Toxin #6 detected from all sites over the sampling perio d Error bars represent sample standard devi ation.

PAGE 87

87 CHAPTER 4 DISCUSSION The results of this study show that Lyngbya dominated mats in two spring fed rivers in Florida are a potential source of Paralytic Shellfish Toxins ( PSTs ) with toxin profiles similar to Lyngbya mat samples collected and describ ed from Alabama (Carmichael et al. 1997) M orphological examination indicates that the Lyngbya in the mats is L. wollei The 16S rR NA sequence obtained from an isolated filament of Lyngbya shows 95% identity to the L. wollei strain Alabama which also pr oduces PSTs and DNA isolation from a single filament confirmed the presence of genes associated with toxin production, including sxt A and sxt G genes The latter observation provid es additional evidence that L wollei is the source of PSTs observed in Sil ver Glen Springs Lyngbya mats. N ucleotide sequences of sxt A and sxt G genes from isolates showed a high level of identity with the L. wollei Alabama strain, in contrast to lower identities f or sequences from other species. Toxin Extraction Methods utilized to extract PSTs from mat material can influence the identification and quantification of toxins Working with lyophilized material that has been cut or mechanically homogenized allow s for more contact of inner cells with the extract solution. The choice of extracti on solution should focus on avoid ing interconversions of PSTs maintain an acidic pH and allow for extraction of as much of the toxins as possible. The use of acetic acid at 0.1 M concentration is suggested for any fu ture work extracting PSTs from L. wollei The use of Solid Phase Extraction ( SPE ) requires more investigation as C18 SPE did not aid in reducing noise or provide improved response in fluorescence or spectromet ry analyses. An SPE method to concentrate the extract and

PAGE 88

88 provide adde d sensitivity would be helpful for confirmation of suspect toxins such as dcNEO and GTX1. The use of lyophilization alone and in conjunction with C18 did not result in improved sensitivity but instead resulted in complete ion suppression via Liquid Chrom atography/Mass Spectrometry ( LC/MS ) Analytical Methods Analytical m ethods used to detect PSTs extracted from Lyngbya are not all equivalent and may lead to an overestimation of toxicity. The results in this study show significant difference s between meth od s P re chromatographic oxidation Liquid Chromatography/Fluorescence ( LC/FL ) techniques only resulted in peaks corresponding to decarbamoyl toxins dcGTX23 and dcSTX/dcNEO. Samples of Blue Hole Spring (collected 4/2/10) extracted with 0.1 M Acetic acid quantitated via LC/FL resulted in toxin concentrations of 490 g dcGTX2&3 g (dry weight) 1 and 26 g dcSTX/dcNEO g(dry weight) 1 This approach le d to an overestimation of actual toxicity since less toxic variants of PSTs added to the total area of the ox idation peaks. The use of LC/MS provided a more accurate detection of PSTs in samples. LC/MS data of Blue Hole Spring (collected 4/2/10) resulted in concentrations of decarbamoyl toxins of 41 g dcGTX2 g(dry weight) 1 24 g dcGTX3 g(dry weight) 1 and 19 g dcSTX g(dry weight) 1 Taking into account relative toxicity factors, LC/FL analysis of Lyngbya samples yielded 6.4 times higher STX eq uivalents than obtained using LC/MS analyses Decarbamoyl toxins dcGTX and dcSTX were present in every sample, bu t made up less than 13% of the total PST constituents detected (7 % a verage) when based on MS response values. This is similar to percent composition determined for L. wollei mats collected in Alabama ( Table 4 1).

PAGE 89

89 Toxins detected via LC/MS included dcGTX2 &3, dcSTX and L. wollei toxins (LWT s ) 1 6. These are the same toxins reported for mats from the Guntersville Reservoir on the Tennessee River in Alabama (Carmichael et al. 1997, Onodera et al. 1997). Average c oncentrations of toxins (nmol g 1 dry weight) and percent composition in this study and that of Onodera et al. ( 1997 ) are shown in Table 4 1. The levels of dcSTX and dcGTX2&3 were comparable in both studies, but comparisons with L. wollei Toxins cannot be made with the limited quantitative data achi eved in this study. The highest percent composition for PSTs reported in Alabama Lyngbya samples was L WT 2, a toxic derivative of PST. It is unknown which LWT was dominant in this study without the use of standards, but mass spectrometer response was gre atest with LWT 4, a non toxic variant. Toxin Risk Assessment The important question with regard to PSTs is toxicity and potential health threats posed by L. wollei mats Standards of L. wollei toxins 1 6 are needed to facilitate routine analyses of toxi ns to eliminate the need to conduct mouse bioassay (MBA) or some other toxicity screening. With regard to current measures of toxicity, only approximat ions can be made from data acquired utilizing assumptions about relative toxicity f rom previous studies. Table 4 2 shows toxicity data for PSTs detected in L. wollei mats as well as STX. The data reported by Oshima et al. ( 1995 ) and Onodera et al. ( 1997 ) is reported in Mouse Units ( MU ) mol 1 of toxin using the AOAC Method for Paralytic Shellfish Poisons, i.e. 20 g ddY strain male mice One Mouse Unit (MU) is the dose of toxin required to kill a male mouse in 15 minutes via intraperitoneal injection (O nodera et al 1997). Saxitoxin exhibit s the highest toxicity at 2483 MU mol 1 of toxin, based on research conducted by Oshima et al ( 1995 ) Although heavily cited, this

PAGE 90

90 reference is not in agreement with other research, which applies different toxicities to PSTs However, is reported here LWT 1 6 tox icit y data since both were derived using similar techniques Lethal doses of toxic samples of L. wollei collected in Alabama ranged from 150 mg of lyohphilized L. wollei to nontoxic quantities of 1500 mg lyophilized L. wollei per kg mouse with reported saxi toxin equivalents from 0 to 58 g STX eq g(dry weight) 1 In this study, total STX eqs cannot be determined with the limited quantitative data achieved with regard to L. wollei toxins 1 6 but STX eq s based on dcSTX & dcGTX2&3 concentrations represent min imum toxicity ranges from 3 4 to 61 g STX eq /g (dry weight) 1 This indicates L. wollei mats c ollected from Florida springs are comparable in toxicity or more toxic than L. wollei mats from Alabama. Care must be taken when analyzing PSTs originating fr om L. wollei dominated mats in the future. Tradition al methods of detection, such as E LISA and pre column oxidation HP LC/Fluorescence may over or under estimate toxicity values. Future studies assessing PST production in L. wollei should incorporate th e use of postcolumn oxidation LC/FL and LC/MS to determine what toxins are actually present. The use of standards is necessary for the final quantitation of L. wollei toxins and to relate levels to potential toxicity. The utilization of other techniques such as tissue culture bioassay s may also provide additi onal information on actual toxin content of L. wollei in Florida without the need to conduct MBA. Lyngbya wollei is prevalent in many of the Florida springs and is not limited to the springs sample d in this study PSTs have been detected in Lyngbya wollei dominated mats c ollected from multiple Florida s prings in conjunction with work conducted by the

PAGE 91

91 FDEP (Florida Departmen t of Environmental Protection) and the FDOH (Florida Department of Health), including Wakulla Springs, Fanning Springs, Homosassa Springs Alexander Springs Juniper Springs and multiple springs in the Ichetucknee Spring System PSTs are known to bioaccumulate in bivalves through filter feeding or exposure to extracellular toxins but the thick sheath associated with Lyngbya filaments inhibit filter feeding and reduces the leaching of toxins into the surrounding water. When Lyngbya senesces or when toxins are released from the Lyngbya sheath the PSTs are thought to degrade to non toxic derivatives or di lute to a non toxic concentration (Carmichael et al. 1997) PSTs do not, however, degrade quickly in a buffered freshwater system, such as the freshwater springs of Florida, an d may remain intact until photolytic or bacterial degra da tion acts on the toxins (Jones and Negri, 1997). On the other hand, the high flow of the spring systems does allow for rapid dilution of any toxins released There is considerable uncertainty regarding the a ctual risk of human exposure to the PSTs pro duced by L. wollei Ingestion of Lyngbya directly is unlikely and the bioaccumulation or biomagnification of the toxins to herbivores or predators is unknown. As levels of accessible drinking water in Florida and other regions decline, sources of surface water with L. wollei presence may be tapped to support the growing need. This may increase the risk of human exposure to PSTs and provide the incentive to procure workable standards for routine analysis of the PSTs associated with L. wollei M anagement d ecisions may be made to remove or treat L. wollei mats as awareness grows and care must be taken to reduce occupational hazard s related to handling treating or removing the mats Recreational exposure is of concern as well

PAGE 92

92 since direct contact with L. w ollei dominated mats in spring runs is likely to occur PSTs may be limited in their ability to cross the dermal layer, but individuals with lesions or cuts in the skin may be of greater risk. To date, there are no known cases directly linking PSTs from L. wollei dominated mats to symptomatic responses It is unknown if this is due to actual risks or if this is because epidemiological data is limited. Currently, data collected by the FDOH in areas dominated by L. wollei is focused on dermatoxic and gast rointestinal responses, with a lack of focus on neurotoxic responses, which can be mild. Questionnaires and reports would have to be extended to include neurotoxic responses associated with PST exposure, such as numbness, tingling sensations, headaches, n ausea, blurred vision, and muscle incoordination. Epidemiological data of this nature might help to provide a better understanding of risks relate d to PSTs derived from L. wollei in the future.

PAGE 93

93 Table 4 1 Toxin content of L.wollei collected from Alabam a and Florida Toxin Alabama Lyngbya nmol/g Percent Composition Blue H ole Spring nmol/g Silver Glen Springs n mol/g dcSTX 40 1.4% 89 100 dcGTX2 110 3.9% 113 98 dcGTX 3 41 1.5% 64 48 a LWT 1 690 24.6% LWT 2 390 46.7% LWT 3 920 0.7% a LWT 4 20 0.7% LWT 5 500 17.8% a LWT 6 97 3.5% Could not quantify a Non toxic derivatives Table 4 2 Toxicity of PST variants reported in MU/mol and r elative toxicity to STX (STX eq) Toxin a Mouse Units ( MU ) /mol Relative Toxicity a STX 2483 1.00 a dcSTX 1274 0.51 a dcGTX2 1617 0.65 a dcGTX 3 1872 0.75 b LWT 1 <10 b LWT 2 178 0.07 b LWT 3 52 0.02 b LWT 4 <10 b LWT 5 326 0.13 b LWT 6 <10 a Oshima et al. 1995 data b Onodera et al. 1997 da ta

PAGE 94

94 LIST OF REFERENCES Berry, J.P., Gantar, M., Gawley, R.E., Wang, M., and Rein, K.S., 2004. Pharmacology and toxicology of pahayokolide A, a bioactive metabolite from a freshwater species of Lyngbya isolated from the Florida Ev erglades. Comparative Biochemistry & Physiology Part C: Toxicol ogy & Pharmacology 139, 231 238 Botana L. 2008. Seafood and freshwater toxins. CRC Press, Taylor and Francis group,Boca Raton Fl. 2 163 197 Camacho, F.A. and Thacker, R.W., 2006. Amphipod herbivory on the freshwater cyanobacterium Lyngbya wollei: Chemical stimulants and morphological defenses. Limnology & Oceanography 51, 29. Campa s M., Prieto Sim o n, B., and Marty, J.L., 2007. Biosensors to detect marine toxins: Assessing seafood safety. Talanta 72, 884 895. Carmichael, W.W., Evans, W.R., Yin, Q.Q., Bell, P., and Moczydlowski, E., 1997. Evidence for Paralytic Shellfish Poisons in the Freshwater Cyanobacterium Lyngbya wollei (Farlow ex Gomont) comb. nov. Appl.Environ.Microbiol. 63, 3104. Co well, B.C., Botts, P.S., 1994 Factors influencing the distribution, abundance and growth of Lyngbya wollei in central Florida. Aquat. Bot. 49, 1 17 Dell'Aversano, C., Hess, P., and Quilliam, M.A., 2005. Hydrophilic interaction liquid chromatography mass spectrometry for the analysis of paralytic shellfish poisoning (PSP) toxins. Journal of Chromatography A 1081, 190 201. Fitzgerald D J Cunliffe D A Burch M D 1999. Development of Health Alerts for Cyanobacteria and Related toxins in Drinking water i n South Australi a. Environ Toxicol. 14. 203 09. Harrington D., Maddox G., Hicks R.,2010. Florida springs initiative monitoring network report and recognized sources of nitrate. Florida Department of Environmental Protection ; Div ision of Environmental Asses sment and Restoration Bureau of Watershed Restoration. Heffernan, J.B., Liebowitz, D.M., Frazer, T.K., Evans, J.M., and Cohen, M.J., 2010. Algal blooms and the nitrogen enrichment hypothesis in Florida springs: Evidence, alternatives, and adaptive managem ent. Ecological applications : a publication of the Ecological Society of America 20, 816. Indrasena, W.M. and Gill, T.A., 2000. Food Chemistry and Toxicology Thermal Degradation of Partially Purified Paralytic Shellfish Poison Toxins at Different Times, Temperatures, and pH. Journal of food science : an official publication of the Institute of Food Technologists 65, 948.

PAGE 95

95 Jones, G.J. and Negri, A.P., 1997 Persistence and degradation of cyanobacterial paralytic shellfish poisons (PSPs) in freshwaters. Wat er Research 31, 525 533. Joyner, J.J., Litaker, R.W., and Paerl, H.W., 2008. Morphological and Genetic Evidence that the Cyanobacterium Lyngbya wollei (Farlow ex Gomont) Speziale and Dyck Encompasses at Least Two Species. Applied & Environmental Microbiolo gy 74, 3710 3717. Kellmann, R., Michali, T., and Neilan, B., 2008. Identification of a Saxitoxin Biosynthesis Gene with a History of Frequent Horizontal Gene Transfers. Journal of Molecular Evolution 67, 526 538. Lawrence, J.F., Niedzwiadek, B., and Menard C., 2005. Quantitative Determination of Paralytic Shellfish Poisoning Toxins in Shellfish Using Prechromatographic Oxidation and Liquid Chromatography with Fluorescence Detection: Collaborative Study. Journal of AOAC International 88, 1714 1732. Mihali, K. T., Carmichael W. W., Neilan B., 2011. A putative gene cluster from a Lyngbya wollei bloomthat encodes paralytic shellfish toxin biosynthesis. PLoS ONE 6(2): e14657. doi:10.1371/journal.pone.0014657 Mihali, K. T., R. Kellmann, and B. Neilan. 2009. Ch aracterization of the paralytic shellfish toxin biosynthesis gene clusters in Anabaena circinalis AWQC131C and Aphanizomenon sp. NH 5. BMC Biochemistry 10:8. Miller, C. 2006. Ecological and Toxicological Assessment of Lyngbya in Florida Springs. Final Rep ort to Florida Department of Health Aquatic Toxins Program. Onodera H, Satake M, Oshi ma Y, Yasimoto T, Carmichael WW, 1997. New saxitoxin analogues from the freshwater filamentous. cyanobacterium Lyngbya wollei. Nat. Toxins. 5(4). 146 51. Oshima Y, 19 95, Post column derivatization liquid chromatographic method for paralytic shellfish toxins. Journal of AOAC International. 78. 528 532. Ravn H Anthoni U, Christophersen C, Nielsen P, Oshima Y. 1995. Standardized extraction method for paralytic shellf ish toxins in phytoplankton Journal o f Applied Phycology. 7. 589 94. Scientific Opinion of the Panel on Contaminants in the Food Chain on a request from the European. 2009. Commission on Marine Biotoxins in Shellfish Saxitoxin Group. The EFSA Journal 1019. 1 76. Seifert, M., McGregor, G., Eaglesham, G., Wickramasinghe, W., and Shaw, G., 2007. First evidence for the production of cylindrospermopsin and deoxy cylindrospermopsin by the freshwater benthic cyanobacterium, Lyngbya wollei (Fa rlow ex Gomont) Speziale and Dyck. Harmful Algae 6, 73 80.

PAGE 96

96 Sivonen K. and Jones G., 1999. Cyanobacterial toxins. Toxic Cyanobacteria in Water: A Guide to Public Health, Significance, and Monitoring and Management. I. Chorus and J. Bertram (Edn). The World Health Organization. ISBN 0 419 23930 8. E & FN Spon. London, UK, 41 111. Speziale, B.J. and Dyck, L.A., 1992. Lyngbya I nfestations : C omparitive T axonomy of Lyngbya wollei Comb. Nov. (cyanobacteria ). Journal of Phycology 28, 693 706. Stevenson, R.J., Pinowska A., Albert in A., Sickman J. 2007. Ecological condition of algae and nutrients in Florida Springs : The synthesis report. Environmental Assessment Section, Florida Department of Environmental Protection. Stucken K., John U., Cembella A., Murillo A.A., Soto Liebe K ., 2010 The smallest known genomes of multicellular and toxic cyanobacteria: comparison, minimal gene sets for linked traits and the evolutionary implications. PLoS ONE 5(2): e9235. doi:10.1371/journal.pone. 0009235 Tamura, K ., Dudley, J., Nei, M., and Ku mar, S., 2007. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) Software Version 4.0. Mol.Biol.Evol. 24, 1596 1599. Teneva, I., Asparuhova, D., Dzhambazov, B., Mladenov, R., and Schirmer, K., 2003. The freshwater cyanobacterium Lyngbya aerugineo coer ulea produces compounds toxic to mice and to mammalian and fish cells. Environ.Toxicol. 18, 9. Thacker, R.W., McLeod, A.M., and McLeod, S.W., 2005. Herbivore induced saxitoxin production in the freshwater cyanobacterium Lyngbya wollei Algological studies 159, 415 425. Tillett, D. and Neilan, B.A., 2000. Xanthogenate nucleic acid isolation from cultured and environmental cyanobacteria. Journal of Phycology 36, 251 258. Vale, C., Alfonso, A., Vieytes, M.R., Romaris, X.M., Arevalo F., Botana, A.M., and Botan a, L.M., 2008. In Vitro and in Vivo Evaluation of Paralytic Shellfish Poisoning Toxin Potency and the Influence of the pH of Extraction. Analytical Chemistry 80, 1770 1776. Wiese M., D'Agostino P.M., Mihali T.K., Moffitt M.C., Neilan B.A., 2010. Neurotoxic Alkaloids: Saxitoxin and Its Analogs. Marine Drugs. 8(7). 2185 2211. Yin, Q., Carmichael, W.W., and Evans, W.R., 1997. Factors influencing growth and toxin production by cultures of the freshwater cyanobacterium Lyngbya wollei Farlow ex Gomont. J.Appl.Phy col. 9, 55. Y i lmaz, M., Phlips, E.J., Szabo, N.J., and Badylak, S., 2008. A comparative study of Florida strains of Cylindrospermopsis and Aphanizomenon for cylindrospermopsin production. Toxicon 51, 130 139.

PAGE 97

97 BIOGRAPHICAL SKETCH Amanda Foss was born and raised in Grand Rapids MI. She earned a B.S. in Environmental Chemistry and Biology from Lake Superior State University in 2006 She was recruited to work for the private company Gre enWater Laboratories in Palatka, Florida analyzing toxin s produced by c yanobacteria. While working at GreenWater, s he received her M.S. in Fisheries and Aquatic Sciences from the University of Florida in the spring of 2011. Upon completion of the program, Amanda continued to work for GreenWater Laboratories isolating and de tecting naturally derived toxins.