Saxitoxins are produced by species of marine eukaryotic dinoflagellates within the genera Alexandrium, Gymnodinium and Pyrodinium as well as by cyanobacteria within a range of species and strains belonging to the Nostocales, that is, Dolicbospermum (Anabaena). (Humpage et al., 1994; Velzeboer et ah, 2000), Apbanizomenon (Ikawa et ah, 1982; Sasner et ah, 1984; Pereira et ah, 2000; Dias et ah, 2002) and Rapbidiopsis (Cylindrospermopsis) raciborskii mainly in Brazil (Lagos et ah, 1999; Molica et ah, 2002) and Scytonema (Smith et ah, 2011) and Oscillatoriales such as Planktotbrix and Microseira (Lyngbya) wollei (Carmichael et ah, 1997; Onodera et ah, 1997). From lakes and reservoirs of the southern USA, Microseira wollei is known to overwinter in the form of benthic mats and rises to form surface mats during the warmer months (Carmichael et ah, 1997).
Cyanobium sp. CENA 142 and Oxynema sp. CENA 135 were among 135 strains isolated from cyanobacteria collected from Cardoso Island and Bertioga mangroves for which both molecular analyses and ELISA showed STXs production (Silva et ah, 2014). For further details, see the review by Testai et ah (2016) and Cires and Ballot (2016).
The production of different STX congeners seems to be strain-specific. Indeed, Cl, C2, GTX2 and GTX3 were found as predominant congeners in environmental samples and isolated strains of Dolicbospermum circinale in Australia, although a hitherto unique toxin composition (exclusively STX and GTX5) was found in a geographically isolated strain from the southwest coast of Australia (Velzeboer et ah, 2000). Ferreira et ah (2001) found that two Apbanizomenon flosaquae strains and samples of a bloom from a reservoir in Portugal contained a specific STX mixture: GTX4 was the dominant analogue, followed by GTX1 and GTX3. A. flosaquae strains in a Chinese lake produced neoSTX, dcSTX and dcGTX3, showing a different toxin profile (Liu et ah, 2006a; Liu et ah, 2006b), whereas A. gracile strains detected in two German lakes produced GTX5, STX, dcSTX and neoSTX (Ballot et ah, 2010).
In Brazilian freshwaters, STXs are attributed to R. raciborskii. In strains isolated from two reservoirs, the contents of total STXs were similar to those reported in D. circinale in Australia (Flumpage et ah, 1994; Lagos et ah, 1999). One of the Brazilian strains showed a toxin profile very similar to that of A. flosaquae, while the other produced only STX and GTX2/3. Other toxin profiles were described for the Tabocas Reservoir in Caruaru (NE Brazil) affected by a R. raciborskii bloom; several STX analogues
(STX, GTX6, dcSTX, neoSTX and dcneoSTX) were identified but no cylindrospermopsin was detected (Molica et al., 2002). Again in Brazil, Castro et al. (2004) reported a R. raciborskii strain isolated from a bloom which contained STX concentrations around 0.3 mg/g DW, which is 4- to 8-fold higher than those of GTX2 and GTX3.
In Lyn. wollei strains isolated from a reservoir in southern USA, GTX2 and GTX3 represented the major STX congeners, whereas STX and neoSTX were not detected (Carmichael et al., 1997).
Planktothrix sp. FP1 has been associated with the production of STXs in a lake in Italy, confirmed in the isolated culture; the toxin profile of this strain included STX, GTX2 and GTX3 (Pomati et al., 2000).
Few data have been published on the cellular contents of STXs in different cyanobacteria. Llewellyn et al. (2001) have reported STX cell quota up to slightly more than 450 ng /106 cells (i.e., 0.45 pg/cell) in a D. circinale strain isolated from an Australian waterbody. Hoeger et al. (2005) estimated the cell quota to be 0.12 pg STXs/cell in D. circinale. Higher cell quotas (up to 1300 fg/cell) are reported for a strain of Scytonema sp. which, however, has very large cells, and in relation to its biomass, with 119 pg/g dry weight the toxin content of this strain was not exceptionally high (Smith et al., 2011). Cell quota up to 0.034 pg/cell of STXeq. were reported in an Apbanizomenon sp. (strain LMECYA 31); in the same culture, very high levels of dissolved STXs were observed in the culture media, especially in the late growth phase, very likely as a consequence of cell lysis and leakage (Dias et al., 2002).
Tables 2.6 and 2.7 give examples of the STX contents of strains and environmental samples, respectively. For further details, see reviews by Funari and Testai (2008), Pearson et al. (2016) and Testai et al. (2016).
Biosynthesis and Regulation
The saxitoxin biosynthesis gene cluster (sxt) was first characterised in Cyl. raciborskii T3 by Kellmann et al. (2008); other characterisations followed from other strains, namely, Dolichospermum circinale AWQC131C, Apbanizomenon sp. NH-5 (Mihali et al., 2009), Rapbidiopisis brookii D9 (Stiiken et al., 2011) and Lyngbya wollei (Mihali et al., 2011). All five sxt clusters encoded biosynthetic enzymes (sxtA, sxtG, sxtB, sxtD, sxtS, sxtU, sxtH/T and sxtl which appear to have diverse catalytic functions) plus regulatory genes (sxtL, sxtN and sxtX) and transporters (Kellmann et al., 2008; Pearson et al., 2010).
Different biosynthetic pathways have been proposed, the most recent by D’Agostino et al. (2014) and reviewed by Pearson et al. (2016), starting with the methylation of acetyl-CoA catalysed by SxtA, followed by a condensation reaction with arginine. Further, the aminotransferase SxtG catalyses the addition of the amidino group from a second arginine residue. The following reactions are cyclisation and desaturation leading to the tricyclic core structure, resulting in decarbamoyl STX (dcSTX). Finally, a carbamoyl group is added to dcSTX by the carbamoyltransferase Sxtl, resulting in the finalised STX molecule.
The N-sulfotransferase (SxtSUL) can modify STX, GTX2 and GTX3, into GTX5-6, C-l and C-2, by transferring a sulphate residue from PAPS (З'-phosphoadenosine 5'-phosphosulphate) to the carbamoyl group. SxtDIOX is proposed to catalyse the Cll hydroxylation of STX followed by subsequent O-sulphation by SxtSUL for biosynthesis of GTX1-4. A combination of sulphation by SxtSUL and SxtN then leads to biosynthesis of the disulphated C-toxins.
STX congeners are mainly produced during late exponential growth phase in laboratory culture (Neilan et al., 2008). The characterisation of the sxt cluster in several genera has enabled the study of molecular mechanisms underlying regulation, based on the identification of the genes sxtY, sxtZ and ompR putatively involved in regulating the sxt cluster, adjacent to the Rapbidiopsis raciborskii T3 sxt cluster (Kellmann et al., 2008). However, so far the direct involvement of the regulatory cluster on STX biosynthesis has not been experimentally demonstrated (Pearson et al., 2016).
Regarding the impact of environmental factors, the analysis of data from Australian field samples suggests that STX production is influenced by environmental factors, particularly alkalinity (pH>8.5), very high ammonia concentration (>1 mg/L) and high conductivity (Neilan et al., 2008). Data from laboratory culture studies further indicate that temperature, culture age, light, pH, salinity and nutrient concentrations affect STX production, although causing a variation of only a 2-4-fold (Sivonen & Jones, 1999; Pearson et al., 2016). However, the impact of a particular environmental modulator strictly depends on strains. As an example, toxin production doubled at higher-than-optimal temperatures with Apbanizomenon sp. LMECYA 31 (Dias et al., 2002), but in contrast to this, an increase in toxin content was observed in Apbanizomenon gracile UAM 529 (Casero et al., 2014) and R. raciborskii CIO (Castro et al., 2004) in response to lower- than-optimal temperature.