The paralytic shellfish toxins (PSTs) are a group of naturally occurring neurotoxic alkaloids. Saxitoxin (STX) is the most researched PST to date, and since its discovery in 1957 [1], 57 analogs have been described. The PSTs are primarily produced in detrimental concentrations during harmful algal bloom (HAB) events [2–5] Over the last few decades, HABs have become more frequent, intense, and span a wider global distribution, the cause of which is still under debate [3,6]. The PSTs can be broadly characterized as hydrophilic or hydrophobic, and can be divided into subgroups based on substituent side chains such as carbamate, sulfate, hydroxyl, hydroxybenzoate, or acetate. Each moiety then imparts a varying level of toxicity [7].

In marine environments, PSTs are primarily produced by the eukaryotic dinoflagellates, belonging to the genera Alexandrium, Gymnodinium and Pyrodinium [8–10]. The toxins are passed through the marine food web via vector organisms, which accumulate the toxins by feeding on PST producing dinoflagellates without apparent harm to themselves [11,12]. These include filter feeding invertebrates such as shellfish, crustaceans, molluscs and also other, non-traditional vectors such as gastropods and planktivorous fish [13]. In freshwater environments the PSTs are produced by prokaryotic cyanobacteria belonging to the genera Anabaena, Cylindrospermopsis, Aphanizomenon, Planktothrix and Lyngbya. Cyanobacterial PST producing blooms result in the contamination of drinking and recreational water resources. In the past, high levels of toxins have been detected in the freshwater resources of many countries such as Australia, Brazil, USA, Mexico, Germany and China [14–22].

Intoxication with PSTs may result in the severe and occasionally fatal illness known as paralytic shellfish poisoning (PSP) or saxitoxin pufferfish poisoning (SPFP) [23–27]. This illness is caused when PSTs reversibly bind voltage-gated Na⁺ channels in an equimolar ratio. This is mediated by the interaction between the positively charged guanidinium groups of STX with negatively charged carboxyl groups at site 1 of the Na⁺ channel, thereby blocking the pore (Figure 1) [28–30]. Currently, there is no antidote for PSP with artificial respiration and fluid therapy the only treatment available. A recent case of PSP involved the death of two fishermen after consumption of the filter feeder bi-valve Aulacomya ater in the Chilean Patagonian Fjords [26]. The threat of PSP is not only a major cause of concern for public health but is also detrimental to the economy. Outbreaks of PSTs often result in the death of marine life and livestock, the closure of contaminated fisheries, while the continual expenditure required for the maintenance and running of monitoring programs, all combine to present a major economic burden around the world [31,32].

This review will focus on the structural diversity of PSTs characterized to date and the biosynthetic and metabolic basis for this diversity. The saxitoxin biosynthetic gene cluster (sxt) was recently identified in cyanobacteria, which now provides insight into the biosynthesis of STX and its analogs [33,34]. A specific suite of analogs can be isolated from a single PST-producing organism, which is directly a result of the evolution of genes present within the organism’s genome [14,33–37]. Naturally occurring PSTs can also be precursors for extracellular metabolic or chemical transformations into new analogs. Knowledge of these transformations may have important implications for the detection, toxicity and removal of PSTs from a contaminated source. Other medicinal uses for PSTs may become more established by screening the bioactivity of less toxic analogs, since their use as a potential local anesthetic has long been known [38,39]. The characterization of PST biosynthesis genes and their potential use in combinatorial biosynthesis, together with the constant discovery of novel analogs (either natural or transformed), is likely to expand the possibilities for the pharmaceutical use of PSTs [40,41].

STX is one of the most potent natural neurotoxins known. A dose of approximately 1 mg of the toxin from a single serving of contaminated shellfish is fatal to humans. STX was the first PST isolated in pure form from the Alaskan butter clam, Saxidomus gigangteus in 1957 [1]. Its highly polar characteristics represent poor conditions for crystallization and hampered structure elucidations for 18 years, until the crystal structure was solved by two groups independently in 1975 [42,43]. STX is an alkaloid with the molecular formula C₁₀H₁₇N₇O₄ (Molecular Weight = 299) and is composed of a 3,4-propinoperhydropurine tricyclic system. STX belongs to the large family of guanidinium-containing marine natural products, due to the presence of two guanidino groups which are responsible for its high polarity [44,45]. Since its initial discovery, 57 naturally occurring STX analogs have been identified in a number of organisms, collectively referred to as the PSTs (Table 1).

Usually a PST- producing organism synthesizes a characteristic suite of toxins made up of several PST analogs. These analogs differ in side group moieties and thus are commonly grouped according to these variable residues. The most commonly occurring PSTs are hydrophilic and have been studied in depth [7]. They may be non-sulfated, such as STX and neosaxitoxin (neoSTX), mono-sulfated, such as the gonyautoxins (GTXs 1–6), or di-sulfated (C1-4 toxins) [7,90]. In addition, decarbamoyl variants of these analogs also exist, including decarbamoyl-saxitoxins (dcSTX, dcneoSTX), decarbamoyl-gonyautoxins (dcGTXs 1–4), and the 13-deoxy-decarbamoyl derivatives (doSTX, doGTX 2,3). Three structural families of SXT are classified by the identity of the R₄ side chain as either N-sulfocarbamoyl, decarbamoyl, or carbamoyl, each with increasing toxicity in mammalian bioassays (Table 2) [7,9,90]. Recently, an increase in screening efforts, coupled with improved methods for detection and structure elucidation, has seen an increase in the number of new PSTs reported in the literature.

A novel group of PSTs with a hydrophobic side chain were identified within the cyanobacterium Lyngbya wollei and are characterized by the presence of an acetate at C13 (LWTX 1–3,5,6) and a carbinol at C12 (LWTX 2,3,5) in place of a hydrated ketone [82]. This was the first report of STX derivatives with a hydrophobic substituent and these toxins have only been found exclusively in the freshwater environment [14,82]. The presence of an acetate side chain in the LWTXs correlated with a decrease in mouse toxicity, while the reduction at C12 resulted in a complete loss of mouse toxicity [82].

Interestingly, Negri et al. reported a novel subclass of analogs containing a hydrophobic R₄ side chain designated GC1-3. These were first isolated and structurally characterized from Australian isolates of the dinoflagellate Gymnodinium catenatum and since have also been identified within Alexandrium catenatum globally [72]. High-resolution mass-spectrometry (MS) and nuclear magnetic resonance spectroscopy (NMR) revealed that GC3 is a 4-hydroxybenzoate ester derivative of dcSTX, while GC1 and GC2 are epimeric 11-hydroxysulfate derivatives of GC3 [83,91]. Negri et al. emphasized that the lipophilic nature of these toxins may lead to an increased potential to bioaccumulate in marine organisms [72]. These novel analogs have also been shown to bind strongly to the voltage gated Na⁺ channel. The binding affinity of GC3 resembles the affinity of the GTXs, whereas the epimer pair GC1 and GC2 bind with a similar affinity compared to the C-toxins [72,92]. More recently, other GC PST analogs have been identified, such as GC4-6, the di-hydroxylated benzoate GC analogs GC1-6a and the sulfated benzoate analogs GC1-6b for which only putative structures have been determined via mass spectrometry (MS) [85]. Due to their hydrophobic nature, these toxins easily escape conventional chromatography methods. The frequently used C18 solid-phase separation is based on polarity and thus hydrophobic compounds are retained on the column and cannot be detected. This is significant from a shellfish monitoring and public safety viewpoint, and presents a major challenge to water authorities [72,93,94].

Recently, Vale et al. reported the isolation of four unusual compounds (denoted A–D) and categorized them as novel STX analogs based on fluorescence emission, ultraviolet absorption maxima and cross-reactivity to a commercial antibody towards STX [86]. These extracts originated from shellfish samples (Semele proficua and Senilia senilis) collected from Luanda and Mussulo Bay, Angola. Compounds A and D were classified as non-N1-hydroxyl PST analogs and compound B as a N1-hydroxyl analog. Even though the presence of G. catenatum and Pyrodinium bahamense has been reported from the coast of Angola, none of the 18 PSTs commonly found in dinoflagellates were identified in these extracts. The authors therefore suggested a possible cyanobacterial source, though neither a definitive chemical structure, nor a PST-producing organism were conclusively identified [86]. Further analysis of the compounds by MS and NMR is required to elucidate these structures and confirm them as STX analogs.

The most exotic STX isolate identified to date was isolated from the Panamanian golden frog Atelopus zeteki and designated zetekitoxin AB (Tables 1 and 2). Zetekitoxin AB was confirmed to be a PST containing a unique 1,2-oxazolidine ring-fused lactam. The binding affinity of zetekitoxin AB for brain, heart, and muscle Na⁺ channels was extremely potent, displaying a toxicity of approximately 580-, 160- and 63-fold greater than STX against each channel, respectively [89].

The constant discovery of novel and diverse STX analogs is a challenge to PST identification and monitoring. Improvement of detection methods will no doubt uncover new natural forms of STX, however, we are still only beginning to understand the mechanisms by which these complex molecules are produced in nature.

Naturally occurring PSTs may be structurally modified by various biological factors. In some cases, these biotransformations can result in new PSTs that cannot be biosynthesized by cyanobacteria or dinoflagellates alone (Figure 2). In addition, less toxic PSTs may be converted into analogs with greater toxicity (e.g., C-toxins→GTXs) or vice versa. Therefore, a clearer understanding of PST biotransformation is needed for predicting more accurate levels of toxicity. This knowledge may also allow for a mechanism of detoxification to be established and utilized in the water supply and shellfish farming industries.

Cell extracts of PST-producing dinoflagellates are capable of enzymatically modifying PSTs. Oshima et al. demonstrated that GTX2 + 3 can be converted into GTX1 + 4 by incubation with Alexandrium tamarense homogenate [92]. Introduction of a sulfate moiety on the carbamoyl group, resultingin the formation of C1 and C2 toxins, has been shown following incubation with G. catenatum homogenate [44,99]. In these organisms, biotransformation is likely to occur via inherent STX tailoring enzymes which are a part of the SXT biosynthetic pathway encoded within the organism.

Due to differences in the toxin profiles of filter-feeding invertebrate PST vectors and causative producing organisms, various studies have been conducted to monitor toxin biotransformation [84,100–105]. Enzymatic transformation of carbamoyl and carbamoyl-N sulfated toxins into the decarbamoyl compounds was detected within the little neck clam, Prothotheca staminea [106]. In addition, the conversion of the GTXs and neoSTX to STX by reduction of the O22-sulfate and N1-hydroxyl groups, respectively, has been observed within the homogenate of the scallop Placopecten magellanicus [107].

GC1-3 can be converted into dcSTX, as has been confirmed in vitro through incubation of semi-purified GC toxins with bivalve digestive glands [93]. Similarly, the recently identified M-toxins (M1-5) are reportedly bivalve metabolites of the PSTs and are not present in PST- producing microalgae [56]. The M-toxins constitute an important toxin fraction in mussels contaminated by A. tamarense and G. catenatum and have been detected in shellfish, including mussels, cockles and clams [56,86]. These findings are similar to previous reports on the isolation of 11-saxitoxinethanoic acid (SEA), a novel PST from the xanthid crab Atergatis floridus, inhabiting the pacific coast of Shikoku Island [87]. Other examples include a novel carbamoyl-N-methylsaxitoxin (STX-uk) isolated from the Bangladeshi freshwater puffer Tetraodon cutcutia [88]. These exotic STX analogs are likely products of toxin transforming enzymes within the vector organism or its associated microorganisms. However, the mechanism of enzymatic transformation in these organisms is yet to be elucidated [56,86–88,106–109].

Biotransformation of the PSTs by bacteria was first suggested many years ago by Kotaki et al., who proposed that marine bacteria, such as Vibrio and Pseudomonas spp., are capable of metabolizing PSTs [110]. In addition, isolates from the viscera of marine crabs, snails and the marine red algae Jania sp., were studied and demonstrated transformation GTX derivatives into STX through reductive eliminations [110,111]. Bacterial conversion of GTX1-4 to STX and neoSTX is reportedly due to the bacterial thiol compounds glutathione and 2-mercaptoethanol [112]. The ability of bacteria to degrade PSTs has been further described by Smith et al., who screened marine bacterial isolates from various shellfish species for their ability to metabolize a range of PSTs, such as GTX1-5, STX and neoSTX, suggesting that bacteria might play an important role in the clearance of PSTs from bivalve molluscs [113]. Novel strains of Pseudoalteromonas haloplanktis, isolated from the digestive tracts of blue mussels (Mytilus edulis) have been reported to possess the ability to reduce the overall toxicity of a PST mixture of algal extracts by 90% within three days [114,115]. Catabolism of the PSTs most likely occurred via oxidation reactions catalyzed by oxidases and peroxidases into aliphatic products for subsequent use in purine and arginine metabolism, although this is speculated, as no catabolized PST products could be identified [115]. Degradation has also been observed during the passage through a bioactive treatment plant, leading to a decrease in predominant C-toxins and an increase of GTX2 + 3 which display relatively higher toxicity [116].

The structure of STX has been known for 53 years and the discovery of novel STX analogs has continued steadily ever since. Today, 57 PST analogs have been reported. With more sensitive detection methods, new STX analogs will most likely continue to be identified, with new functional moieties and possibly novel bioactivity. Despite extended research on the role of saxitoxin and its analogs as a sodium channel blocker, the effect of these toxins on the environment, and the genes that are responsible for their production, there is still a vast gap in knowledge in regards to their potential intracellular role within the producing organism. Nevertheless, it is possible that the different analogs display varying functions within the cells due to their partial differences in charges and chemical properties. More studies are needed to elucidate the localization of saxitoxin and its derivatives might provide clues to the potential role of the PST analogs within the producing organism. In the future, a better understanding of the intracellular and extracellular functions of STX might open more avenues for pharmaceutical applications.

Since PSTs are produced by distantly related organisms, spanning two domains, including cyanobacteria, dinoflagellates and the Panamanian golden frog, it is possible that their occurrence in nature is more widespread than we know. Further investigations are needed to elucidate the extent of their distribution, diversity and their fundamental biology, such as their biosynthesis, metabolic and eco-physiological function. This is in addition to the role of chemical transformation of the different toxins in shellfish and the environment.

Future research is also needed to understand the integration of PST biosynthesis within the overall cell metabolism and the possible recruitment of enzymes from other biosynthetic pathways for PST bioconversions. Proteomic and transcriptomic studies are likely to provide a link between STX biosynthesis, regulation and cellular metabolism. It is expected that data will allow us to acquire a better understanding of the conservation of the SXT biosynthesis pathway at the enzymatic level in comparison to the genetic level, may give further insight into the molecular function of these toxins and also lead to clues of their evolutionary history. In future, characterization of PST biosynthetic genes from dinoflagellates and comparison with cyanobacterial genes will also aid in our understanding of the evolutionary history of these genes with regard to their origin and transfer.

PSP is a serious health problem and its incidence has continued to rise on a global scale. PSTs negatively impact the fisheries industry globally and the development of novel methods of detoxification is essential from a human health and financial perspective [104,113,162]. The enzymatic basis for the structural diversity of PSTs is now beginning to be understood from the genetics of their biosynthesis in cyanobacteria and characterization of transformations catalyzed by bacteria, marine invertebrates and mammals. Biotransformation pathways could also be manipulated to efficiently remove toxins from water supplies. Specific enzymes or bacterial strains that degrade PSTs could be introduced into shellfish to assist detoxification. Currently, the PSTs represent extraordinary potential for pharmacy. This potential is likely to increase as we continue to gain a better molecular understanding of the PSTs, leading to future prospects of their use in combinatorial biosynthesis for the production of novel alkaloids with beneficial application.