Natural and Derivative Brevetoxins: Historical Background, Multiplicity, and Effects

1Florida red tide inhalation toxicity has been documented in the Western literature since 1844. Earlier accounts from Spanish explorers of an “irritating essence” predate 1600. Past explanations for Florida red tide respiratory distress have included World War I nerve gas release, “fast”- and “slow”-acting toxins, a phosphorylated organic molecule, and green particles as fragments of a marine alga. Descriptions of Florida red tide complexity have mirrored increases in capability of analytical detection. Before 1970, the description of red tides was based on environmental observations of irritating aerosols, fish-killing substances, and hemolytic factors [reviewed in, ]. Because such small amounts of bioactive material were necessary to elicit the toxicologic responses, much of the initial pharmacology/toxicology was understood long before toxin structures were known. Investigators in the early 1970s, using the new field of high-performance liquid chromatography, purified two toxins with very similar spectroscopic characteristics, their differences being the presence of either an exomethylene-conjugated aldehyde or the corresponding allylic alcohol. These two materials of “interchangeable nature” (T46 and T47, or T17 and T34) later became known as the polyether brevetoxins PbTx-3 and PbTx-2, respectively. The structure of PbTx-2 (brevetoxin B) was determined by Nakanishi’s group in 1981 to be a trans-fused polyether ladder toxin with syn relative stereochemistry across each side of the molecule. Additional toxins were structurally characterized by comparison with the PbTx-2 spectroscopic data; PbTx-3 was proposed as the reduced form of PbTx-2 by and Chou and colleagues. Subsequent reduction of PbTx-2 with tritium-labeled sodium borohydride allowed Baden and colleagues to produce isotopically labeled brevetoxin with sufficient specific activity for radioimmunoassays and later to characterize the brevetoxin binding site in rat brain synaptosome sodium channels.

2described brevetoxin A, possessing a slightly different polyether backbone, from cultures of Karenia brevis. This molecule, denoted PbTx-1, is very similar to the PbTx-2 structure in that both possess lactone functionality in the A-ring and a series of relatively rigid rings that form a ladder structure, and both terminate in an identical, very reactive α, β-unsaturated aldehyde side chain. It is noteworthy that PbTx-1 is the only natural toxin known to possess rings with five, six, seven, eight, and nine members, all in the same molecule. Seven additional structures followed in quick succession (Figure 1), the structures being determined by comparison of spectra with those of the two parent toxins, PbTx-1 and PbTx-2.

3By molecular modeling,, ) and determined that the structures of both PbTx-1 and PbTx-2 are relatively linear with a bend approximately mid-molecule, possess a lactone functionality in the A-ring, have a strictly rigid region in the terminal four rings, possess a side chain allowing modest modification at the molecules’ termini, and have a spacer region that separates the rigid region from the A-ring lactone. All natural brevetoxins and synthetic derivatives with full activity possess all these features. Numerous publications describe many of the salient features of the brevetoxin molecules and cite the modulating effects of altering the lactone functionality, of changing the nature of the side chain, of inducing additional flexibility in the middle of the molecule, or combinations thereof [e.g.,; ]. All derivatives of natural toxin examined to date reduce toxicity; that is, no derivative has been produced synthetically that is more toxic than PbTx-1 or PbTx-2, the toxins reputed to be the parent molecules. In fact, it can be argued that all other natural toxins are produced biosynthetically by K. brevis using the two parent backbones as precursors.

4had produced totally synthetic brevetoxin PbTx-2, which had receptor binding properties and spectroscopy identical to the natural material, confirming that the structure proposed by was indeed correct. Subsequent synthesis of the brevetoxin A molecule (PbTx-1) was likewise accomplished. The production of truncated brevetoxin by Nicoloau and colleagues, a compound possessing all of the “required” features of intact brevetoxin except for the B–E ring spacer region, was essentially inactive in patch-clamp experiments, although it had some affinity for receptor site 5 on the sodium channel.

5As a result of the multiplicity of effects demonstrated by this combination of metabolically relevant structural modifications, receptor binding pharmacology, and single-channel kinetic studies, it was postulated that Florida red tide potency in situ would be a complex phenomenon and the toxicologic consequences of red tides would be based on the amounts and activities of the brevetoxins present.

6The purpose of this mini-monograph introduction is to describe the complex nature of exposure to Florida red tide toxins, as reference for articles presented later in this monograph (Figure 2). Although most of the studies have been conducted with PbTx-3 (a natural toxin possessing all of the requisite structural characteristics for full activity), the investigators acknowledge that many studies will require repetition for toxins missing one or more activity loci, or for complex mixtures that may include metabolites and antagonists in addition to toxin. Presumably, any bioactive agents produced by the toxigenic organism or any metabolites free in the water have the potential to become aerosolized. Further, we believe that the overall signs and symptoms observed in humans will depend on the particular combination of activities elicited by the toxins and derivatives/metabolites.

7We restrict ourselves to a discussion of those toxins, antagonists, and metabolites that have been described from seawater or culture medium, although we realize that aerosols may contain cell particles and organelles as well as bacteria, fungi, and other organic materials in the mixed microlayer of the coastal ocean. The metabolites produced by shellfish, fish, and other marine creatures, although extremely important for seafood poisonings, are not considered to be relevant to inhalation toxicology, and they have not been detected by us in air filters.

8Figure 3 illustrates the model “brevetoxin” that encompasses all the salient features of toxins that are “fully active” at their binding site, voltage-sensitive sodium channels (VSSCs; membrane potential-gated ion channels). Modeling work initiated by described a brevetoxin PbTx-2 molecule believed to be composed of three distinct segments: a relatively rigid fourring H–K ring system, an A-ring lactone, and a B–G ring region with limited flexibility. Homologous ring systems exist in the brevetoxin A backbone. All active toxins have these essential characteristics whether they are members of the brevetoxin B backbone or the brevetoxin A backbone (Figure 1).

9Active toxins bind to site 5 on the α-subunit of the VSSC and are thought to orient “head-down” into the channel, intercalating between the α-helices of domains III and IV of the VSSC (Figure 4).

10The α-subunit is a single polypeptide glycoprotein, which possesses a 4-fold homology. Each homologous domain contains six trans-membrane α-helices (S1–S6), which alternately extend through the membrane, with short-to-long trans-helical peptide regions. Three of the helices in each domain are neutral in charge; two of the helices are hydrophobic, and one helix in each domain is highly positively charged. The highly charged S4 helix in each domain is thought to respond allosterically to changes in membrane potential, a “voltage sensor.” Allosteric realignment of all four S4 helices results in a change in channel configuration from “closed” (C), to “open” or conducting (O). Channels inactivate or become nonconducting (I) by internal insertion of the isoleucine–phenylalanine–methionine tripeptide region of the intra-cellular polypeptide connecting domains III and IV into the ion-conducting central pore. Normally, VSSCs open in response to membrane depolarization and subsequently inactivate in the late phase after activation. Channels return to the allosterically closed configuration during membrane repolarization. All natural brevetoxins have four distinct activities that alter the normal C→ O→ I triad of VSSCs: a) the activation potential is shifted to more negative potentials, favoring a C→O allosteric change at normal resting potential; b) a longer mean open time O, which can be influenced by a number factors, but which results in the channel being in the open configuration longer; c) an induction of sodium ion subconductance states in addition to the normal 21 pS rate; and d) an inhibition of inactivation O→I. Derivatives of natural toxin produced in the laboratory, as well as metabolites, knock out one or more of the observed activities in single ion channel patch-clamp experiments. The resulting overall consequences of exposure to multiple natural toxins and metabolites are complex and difficult to quantify mechanistically. How metabolites or modifications to toxin loci modulate activity is summarized in Table 1.

11PbTx-11, PbTx-12, and PbTx-tbm (brevetoxin PbTx-2 lacking the side chain tail) are new toxins found in cultures and in the field (;; Bourdelais et al., unpublished results). Their toxicology is almost completely unknown except that they competitively bind at site 5 on the sodium channel. Figure 6 shows the chemical structure of brevenal, the antagonistic material isolated from both laboratory cultures and bloom waters (, in press). Brevenal derivatives have also been detected in our cultures. Brevenal resembles hemibrevetoxin, isolated by. Cytotoxicity was reported for hemibrevetoxin, but no antagonistic activity was recorded.

12From the data collected, it is important to note again that not only are there two structural backbones for the brevetoxin molecules, but there are at least 13 different derivatives thereof, and each of these derivatives possesses a specific toxicity that is correlated to its binding affinity on VSSCs. However, from past work it is clear that derivatives produced at specific loci on the brevetoxin backbone impart a differential set of “activities” that encompass effects on activation potential, mean open time, inactivation, and persistent depolarization due to subconductance state prevalence.

13Many of the examined derivatives can be produced by intermediary metabolism and environmental decomposition, suggesting that environmental episodes may exhibit great variation in both degree and effect. The added presence of the brevenals would tend to moderate the overall consequences. Its waxing and waning presence during a red tide may provide a partial explanation for anecdotal accounts of high toxigenic organism cell counts in the absence of pronounced fish kills or respiratory irritation, or of relatively low cell counts and reports of “respiratory irritation”.

14What is there to say, then, about the overall pulmonary effects of these polyether natural products? First, in 1947 Woodcock collected droplets of vapor from active red tides and from normal seawater. The droplets from the red tide areas differed significantly from those from non-red tide areas. Red tide aerosol droplets contained small, greenish granules similar in color and shape to those observed in the red tide organism (K. brevis). The granules, which were approximately 1–2 μm in diameter, are an ideal size to become delivery vehicles for pulmonary exposure. Evident from Woodcock’s work is not only that respiratory irritants become airborne from red tides but also that actual particles of the organism become airborne as well. Were these particles to be the delivery mechanism for at least part of the inhaled red tide brevetoxins, then it could be concluded that any toxin or metabolite present in the organism has the potential to be inhaled—perhaps as part of an intracellular storage granule or organelle. The search for the storage depot of brevetoxins in the K. brevis organism has been fruitless. Density centrifugation of whole Karenia cells indicates that the toxins are intracellular. Density centrifugation of cell lysates fails to reveal any concentration of brevetoxins within any subcellular particle (Baden DG and Bishop B, unpublished observations). Nor has immunologic staining of the organism for toxin produced satisfactory results (Naar J, unpublished observations).

15As described by, the brevetoxins have effects at 1,000-fold lower concentrations in pulmonary systems than they have in any neuronal system studied to date. Moreover, the relative concentrations of toxin and antagonist in the pulmonary system are reversed from those exhibited by neuronal systems. That is to say, brevenal is a more potent antagonist in pulmonary systems than it is in neuronal or fish bioassays. This is an interesting set of data because it tends to support the contention that there is something very different about the pulmonary receptor when compared with site 5 of VSSC in the central nervous system. Either way, further exploration of the binding phenomenon in pulmonary tissues is warranted.

16The complex mixtures presented by Florida red tides, all of which are known to bind to a single class of orphan receptors associated with VSSC site 5, provide insight into the neuronal effects of brevetoxins. That there are recognized effects in pulmonary systems and inhibitory effects on cathepsin catabolic enzymes only serves to illustrate the truly complex nature of the toxicology arising from the family of polyether natural products derived from K. brevis. Many of these effects are only now becoming clear. A nagging question is whether the pulmonary receptor is associated with the epithelial sodium channel (a ligand-gated ion channel) or if it is a new orphan receptor. Further, the inhibition of cathepsin within its active site by brevetoxin points to another potential activity for Florida red tide metabolites.

17We conclude that the toxicology surrounding exposure to Florida red tides involves multiple biotoxins and antagonists, and these interact with at least three receptors located in neuronal, pulmonary, and enzymatic regulatory systems of living organisms. The result is a complex combination of acute and chronic signs and symptoms in both animals and humans exposed to aerosolized bioactive materials produced by K. brevis.