Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. 2008 Jul; 74(13): 4044–4053.
Published online 2008 May 16. doi:  10.1128/AEM.00353-08
PMCID: PMC2446512

Biosynthetic Intermediate Analysis and Functional Homology Reveal a Saxitoxin Gene Cluster in Cyanobacteria


Saxitoxin (STX) and its analogues cause the paralytic shellfish poisoning (PSP) syndrome, which afflicts human health and impacts coastal shellfish economies worldwide. PSP toxins are unique alkaloids, being produced by both prokaryotes and eukaryotes. Here we describe a candidate PSP toxin biosynthesis gene cluster (sxt) from Cylindrospermopsis raciborskii T3. The saxitoxin biosynthetic pathway is encoded by more than 35 kb, and comparative sequence analysis assigns 30 catalytic functions to 26 proteins. STX biosynthesis is initiated with arginine, S-adenosylmethionine, and acetate by a new type of polyketide synthase, which can putatively perform a methylation of acetate, and a Claisen condensation reaction between propionate and arginine. Further steps involve enzymes catalyzing three heterocyclizations and various tailoring reactions that result in the numerous isoforms of saxitoxin. In the absence of a gene transfer system in these microorganisms, we have revised the description of the known STX biosynthetic pathway, with in silico functional inferences based on sxt open reading frames combined with liquid chromatography-tandem mass spectrometry analysis of the biosynthetic intermediates. Our results indicate the evolutionary origin for the production of PSP toxins in an ancestral cyanobacterium with genetic contributions from diverse phylogenetic lineages of bacteria and provide a quantum addition to the catalytic collective available for future combinatorial biosyntheses. The distribution of these genes also supports the idea of the involvement of this gene cluster in STX production in various cyanobacteria.

Paralytic shellfish poisoning (PSP) toxins are among the world's most potent and pervasive toxins and are considered a serious toxicological health risk that may affect humans, animals, and ecosystems worldwide (18, 36). These toxins block voltage-gated sodium and calcium channels and prolong the gating of potassium channels (21, 53, 59), preventing the transduction of neuronal signals. It has been estimated that more than 2,000 human cases of PSP occur globally every year at a mortality rate of 15% (16). Moreover, coastal blooms of productive microorganisms result in millions of dollars of economic damage due to PSP toxin contamination of seafood and the continuous requirement for costly biotoxin monitoring programs. Early warning systems to anticipate the occurrence of paralytic shellfish toxin (PST)-producing algal blooms, such as PCR and enzyme-linked immunosorbent assay-based screening, are as yet unavailable due to the lack of data on the genetic basis of PST production.

Saxitoxin (STX) is a tricyclic perhydropurine alkaloid that can be substituted at various positions, leading to more than 30 naturally occurring STX analogues (4, 5, 28, 32, 33, 63). Although STX biosynthesis seems complex and unique, organisms from two kingdoms, including certain species of marine dinoflagellates and freshwater cyanobacteria, are capable of producing these toxins, apparently by the same biosynthetic route (47). In spite of considerable efforts, none of the enzymes or genes involved in the biosynthesis and modification of STX have been previously identified (15, 39, 40, 44, 62). Here, based on previously published knowledge regarding the possible steps in STX biosynthesis (47), together with information from our recent in vitro biosynthesis of STX (22), we used an approach that employed reverse genetics to identify the candidate STX biosynthetic gene cluster (sxt) in the cyanobacterium Cylindrospermopsis raciborskii T3 (23). Since this organism is not genetically transformable, mutagenic characterization of the sxt cluster was not possible. However, here we present the bioinformatically inferred functions for most of the open reading frames (ORFs) in this gene cluster and provide evidence for their role in STX metabolism via liquid chromatography-tandem mass spectrometry (LC-MS-MS) screening of the biosynthetic intermediates in concentrated cell extracts of C. raciborskii T3. The in silico functional assignment of sxt genes and the chemical detection of STX biosynthesis intermediates have enabled a thorough revision of previous knowledge concerning the known STX biosynthetic pathway.


Cyanobacterial cultures.

Cyanobacterial strains used in the present study (Table (Table1)1) were grown in Jaworski medium (55) in static batch culture at 26°C under conditions of continuous illumination (10 μmol m−2 s−1) with fluorescent cool-white light.

Distribution of the sxt genes in toxic and nontoxic cyanobacteriaa

Characterization of the sxt gene cluster.

Total genomic DNA was extracted from cyanobacterial cells by lysozyme/sodium dodecyl sulfate/proteinase K lysis following phenol-chloroform extraction as described previously (29). DNA in the supernatant was precipitated with 2 volumes of −20°C ethanol, washed with 70% ethanol, dissolved in Tris-EDTA buffer (10:1), and stored at −20°C. In a previous study (R. Kellmann, T. K. Mihali, and A. Neilan Brett, submitted for publication), a gene (sxt1) was identified that was exclusively present in PSP-toxic strains of cyanobacteria. It encoded an O-carbamoyltransferase that was predicted to catalyze the formation of the O-carbamoyl side chain of saxitoxin. Adaptor-mediated PCR (49), with modifications as described by Moffitt and Neilan (27), was used to determine the sequence of genes adjacent to sxt1. Short adaptor DNA was ligated to restriction enzyme-digested genomic DNA, and a specific genomic outward-facing primer was then used with an adaptor primer to amplify a region of the genome. Twenty picomoles of T7 adaptor was added to each reaction mixture, containing 1 μg of genomic DNA, 10 U of blunt-ended restriction enzyme, and 5 U of T4 ligase (Promega) in 1× One-Phor-All buffer (Amersham/Pharmacia). The one-step digestion and ligation reaction mixture was incubated at room temperature overnight. The single-stranded end of the adaptor was blocked in a solution containing 1× PCR buffer (Fischer Biotech), 4 mM MgCl2, and 12.5 μM dideoxynucleotide triphosphate with 1 U of Taq DNA polymerase (Fischer Biotech). Thermal cycling was performed using a PCR Sprint temperature cycling system machine (Hybaid Limited) with an initial step at 70°C for 15 min followed by 10 cycles of DNA denaturation at 95°C for 10 s, DNA reannealing at 40°C for 1 min, and extension of the strand with dideoxynucleotide triphosphate at 70°C for 1 min. Following the PCR cycles, the reaction mixture was incubated with 1 U of shrimp alkaline phosphatase (Boehringer Mannheim, Göttingen, Germany) at 37°C for 20 min, and the enzyme was heat inactivated at 85°C for 5 min.

The flanking-region PCR mixture contained 1 to 2 μl of adaptor-ligated DNA, 10 pmol of adaptor primer, and 10 pmol of a genome-specific oligonucleotide primer. Primer sequences are given in the supplemental material. PCR cycling was performed as described above, with DNA strand extension at 72°C for 5 min. The primer annealing temperature was decreased (from 65 to 55°C) by 1°C at each cycle, followed by primer annealing at 55°C for a further 25 cycles. PCR amplicons were separated by agarose gel electrophoresis using TAE buffer (40 mM Tris-acetate, 1 mM EDTA, pH 7.8) and visualized by UV translumination after staining in ethidium bromide (0.5 μg/ml). PCR amplicons were excised and purified for sequencing as described below.

Automated DNA sequencing was performed using a PRISM BigDye cycle sequencing system and a model 373 sequencer (Applied Biosystems). Sequence data were analyzed using ABI Prism Autoassembler software, and percent similarity and identity to other translated sequences were determined using BLAST in conjunction with the National Center for Biotechnology Information (NIH). The sxt gene clusters were assembled using the software Phred, Phrap, and Consed (http://www.phrap.org/phredphrapconsed.html), and ORFs were manually identified. The correctness of the assembly was verified by PCR using genomic DNA as the template, genome-specific primer pairs that targeted overlapping sequence stretches along the entire sxt gene cluster, and sequencing of the obtained PCR amplicons.

MS analysis of STX intermediates.

Extracts were prepared from the cell mass from 1-liter cultures of the PSP-toxic C. raciborskii strain T3 and the non-PSP-toxic C. raciborskii strain AWT205. Cells were lysed in a French press, and the lysate was centrifuged at 100,000 × g for 1 h at 4°C. The supernatant was filtered using a spin filter with a nominal molecular mass cutoff of 10,000 Da. Bacterial extracts and STX standards were analyzed by high-performance liquid chromatography (HPLC; Thermo Finnigan Surveyor and autosampler) coupled to an ion trap mass spectrometer (Thermo Finnigan LCQ Deca XP Plus) fitted with an electrospray source. Separation of analytes was obtained on a Phenomenex Luna 3-μm C18 column (2.1 mm by 150 mm) at 100 ml/min. Analysis was performed using a gradient starting at 95% water -10 mM heptafluorobutyric acid in water (buffer A). This mixture was maintained for 10 min and was then ramped to a 100% organic mixture (acetonitrile [buffer B]) over 30 min. Conditions were maintained using 100% buffer B for 10 min to wash the column, and then the conditions were returned to use of 100% buffer A and again maintained for 10 min to equilibrate the column for the next sample. This resulted in a runtime of 60 min per sample. Sample volumes of 10 to 100 ml were injected for each analysis. The HPLC eluate directly entered the electrospray source, which was programmed as follows: electrospray voltage, 5 kV; sheath gas flow rate, 30 arbitrary units; and auxiliary gas flow rate, 5 arbitrary units. The capillary temperature was 200°C and had a voltage of 47 V. Ion optics were optimized for maximum sensitivity before sample analysis was performed using the autotune function of the instrument with a standard toxin solution. Mass spectra were acquired in the centroid mode over an m/z range of 145 to 650. The mass range was set at “normal,” with 200 ms of maximum ion injection time and automatic gain control engaged. Tandem mass spectra were obtained over an m/z range relevant to the precursor ion. Collision energy was typically 20 to 30 ThermoFinnigan arbitrary units and was optimized for maximal information, using standards where available.

Nucleotide sequence accession number.

Nucleotide sequences were submitted to GenBank and are available under accession number DQ787200.


Identification and sequencing of the sxt gene cluster in C. raciborskii T3.

A biosynthetic pathway has been previously proposed based on the pattern of substrate incorporation into STX, suggesting arginine, acetate (via acetyl-coenzyme A [acetyl-CoA]), and methionine methyl (via S-adenosylmethionine [SAM]) as precursors (47). The first and most unusual of these reactions has been suggested to be a Claisen condensation of arginine and acetate (48). Claisen condensations on amino acids are rare, but do occur, for instance, in the biosynthesis of tetrapyrroles, biotin, and arphamenine AB and in the conversion between glycine, serine, and threonine (2). This reaction is then followed by the transfer of a guanidino group from a second arginine to the first intermediate. The three heterocycles of STX are formed by unknown reactions, and the resulting intermediate is methylated to form a side chain, which is hydroxylated via an epoxide and then O-carbamoylated. Finally, one of the ring carbons (C-12) receives two hydroxy groups (Fig. (Fig.11).

FIG. 1.
Hypothetical biosynthesis pathway of STX as proposed by Shimizu et al. (47). Hypothetical intermediate metabolites are labeled with letters in brackets. The reaction steps are as follows: 1, Claisen condensation reaction between acetate and arginine; ...

This previously proposed pathway for the biosynthesis of STX suggested the involvement of enzymes with the catalytic activities of a class II aminotransferase, amidinotransferase, SAM-dependent methyltransferase, hydroxylases, and O-carbamoyltransferase. Our recent in vitro biosynthesis studies (22) confirmed arginine, acetyl-CoA, SAM, and carbamoylphosphate as precursors for STX biosynthesis, further supporting the involvement of an O-carbamoyltransferase in the biosynthesis of STX. When we screened for putative STX-biosynthetic enzymes in cyanobacterial isolates, a gene encoding an O-carbamoyltransferase in C. raciborskii T3 was initially detected via degenerate PCR (Kellmann et al., submitted). Further investigation showed that homologues of sxtI were exclusively present in PSP toxin-producing strains of four cyanobacterial genera (Table (Table1);1); thus, sxtI represents a good candidate gene for PSP toxin biosynthesis. The sequence of the complete putative STX biosynthetic gene cluster (sxt) was then obtained by genome walking up- and downstream of sxtI in C. raciborskii T3 (Fig. (Fig.2).2). In C. raciborskii T3, this sxt gene cluster spans approximately 35,000 bp, encoding 31 ORFs (Table (Table2).2). The cluster also included other genes encoding previously predicted STX-biosynthesis enzymes, including a methyltransferase (sxtA1), a class II aminotransferase (sxtA4), an amidinotransferase (sxtG), and dioxygenases (sxtH), in addition to the O-carbamoyltransferase (sxtI). PCR screening of selected sxt ORFs in a total of 18 cyanobacterium strains showed that they were exclusively present in PSP toxin-producing isolates (Table (Table1),1), indicating the association of these genes with the toxic phenotype. In the following passages we describe the ORFs in the putative sxt gene cluster and their predicted functions, based on bioinformatic analysis and LC-MS-MS data on biosynthetic intermediates and in vitro biosynthesis, when applicable.

FIG. 2.
Structural organization of the sxt gene cluster from C. raciborskii T3. Abbreviations used are as follows: IS4, insertion sequence 4; ompR, transcriptional regulator of ompR family; hisA, two-component histidine kinase; orf24, ORF 4. The scale indicates ...
The sxt genes from C. raciborskii T3 and their predicted functions

Functional identification of the parent molecule biosynthetic genes.

Bioinformatic analysis of the sxt gene cluster revealed that it contains a previously undescribed example of a polyketide synthase (PKS)-like structure named sxtA. SxtA possesses four catalytic domains, SxtA1 to SxtA4. An iterated PSI-BLAST search revealed low sequence homology of SxtA1 to SAM-dependent methyltransferases. Further analysis revealed the presence of three conserved sequence motifs in SxtA1 (278-ITDMGCGDG-286, 359-DPENILHI-366, and 424-VVNKHGLMIL-433) that are specific for SAM-dependent methyltransferases (19). SxtA2 is related to GCN5-related N-acetyl transferases (GNAT). GNAT catalyze the transfer of acetate from acetyl-CoA to various heteroatoms (31) and have been reported in association with other unconventional PKSs, such as PedI (37), where they load the acyl carrier protein (ACP) with acetate. SxtA3 is related to an ACP and provides a phosphopantetheinyl-attachment site (24). SxtA4 is homologous to class II aminotransferases and was most similar to AONS (8-amino-7-oxononanoate synthase). Class II aminotransferases are a monophyletic group of pyridoxal phosphate-dependent enzymes and are the only enzymes that are known to perform Claisen condensation of amino acids (2). In vitro biosynthesis of STX in C. raciborskii has also been shown to be dependent on the presence of pyridoxal phosphate (22). We therefore reasoned that sxtA performs the first step in STX biosynthesis, involving a Claisen condensation reaction.

It has been proposed that acetate is condensed to arginine in the first step of STX biosynthesis, while the methyl side chain is introduced during a latter step (47). However, the predicted reaction sequence of SxtA, based on its primary structure, is the loading of the ACP (SxtA3) with acetate from acetyl-CoA, followed by the SxtA1-catalyzed methylation of acetyl-ACP, converting it to propionyl-ACP. SxtA4, the class II aminotransferase domain, would then perform a Claisen condensation reaction between propionyl-ACP and arginine (Fig. (Fig.3).3). The putative product of SxtA is thus 4-amino-3-oxo-guanidinoheptane, which is here designated compound A′ (Fig. (Fig.3).3). This sequence of reactions in the first step of STX biosynthesis is inconsistent with that proposed by Shimizu (47) (Fig. (Fig.1).1). To verify this revised pathway for STX biosynthesis based on comparative gene sequence analysis, cell extracts of C. raciborskii T3 were screened by LC-MS-MS for the presence of compound A′ (Fig. (Fig.4)4) and compound B, as predicted by Shimizu (47) (Fig. (Fig.1),1), as well as for arginine and STX as controls. In addition, LC-MS-MS analysis was performed on cell extracts from the non-PSP-toxic strain C. raciborskii AWT205. While arginine and STX were readily detected in extracts from C. raciborskii T3 (Fig. (Fig.4)4) and produced the expected fragment ions (50, 58), there were no measurable levels of compound B. On the other hand, LC-MS-MS data obtained at an m/z of 187 were consistent with the presence of structure A′ from C. raciborskii T3 (Fig. (Fig.4).4). MS-MS spectra showed the expected fragment ion (m/z 170 and 128) (58) after the loss of ammonia and guanidine from structure A′. Arginine was readily detected in cell extracts from C. raciborskii AWT205; however, there were no measurable levels of STX, compound A′, or compound B. LC-MS-MS data strongly supported the predicted function of SxtA and, thus, a revised initiating reaction in the STX biosynthesis pathway.

FIG. 3.
Revised pathway for STX biosynthesis and the putative functions of sxt genes (see text for description).
FIG. 4.
MS-MS spectra of selected ions from cellular extracts of C. raciborskii T3. The predicted fragmentation of ions and the corresponding m/z values are indicated. a, arginine (m/z 175); b, saxitoxin (m/z 300) (refer to Sleno et al. [51] for a more detailed ...

sxtG encodes a putative amidinotransferase, which had the highest amino acid sequence similarity to l-arginine/l-lysine amidinotransferases. We propose that the product of SxtA is the substrate for the amidinotransferase SxtG, which transfers an amidino group from arginine to the α-amino A′ group (Fig. (Fig.3),3), thus producing 4,7-diguanidino-3-oxoheptane (designated compound B′) (Fig. (Fig.3).3). The description of this hypothetical sequence of reactions was also supported by the detection of compound C′ by LC-MS-MS (Fig. (Fig.4).4). Cell extracts from C. raciborskii T3, however, did not contain any measurable levels of B′ (4,7-diguanidino-3-oxoheptane). A likely explanation for the failure to detect the intermediate B′ is its rapid cyclization to form C′ via the action of SxtB. Compound B′ and compound C′ were not detected in extracts from C. raciborskii AWT205.

The sxt gene cluster encodes SxtB, an enzyme similar to the cytidine deaminase-like enzymes from gammaproteobacteria. The catalytic mechanism of cytidine deaminase is a retroaldol cleavage of ammonia from cytidine, which is the same reaction mechanism in the reverse direction as that seen in the formation of the first heterocycle in the conversion from B′ to C′ (Fig. (Fig.3).3). We therefore suggest that SxtB catalyzes this retroaldol-like condensation (see step 4 in Fig. Fig.33).

The incorporation of methionine methyl into STX and its hydroxylation have been studied in detail. Only one methionine methyl-derived hydrogen is retained in STX, and a 1,2-H shift has been observed between acetate-derived C-5 and C-6 of STX. The results of labeled precursor feeding studies indicated that hydroxylation of the methyl side chain of the STX precursor proceeds via epoxidation of a double bond between the SAM-derived methyl group and the acetate-derived C-6 (see steps 5 to 7 in Fig. Fig.1).1). Shimizu (47) proposed that this incorporation pattern resulted from an electrophilic attack of methionine methyl (see step 4 in Fig. Fig.1)1) on the double bond between C-5 and C-6, which would have formed during the preceding cyclization (see step 3 in Fig. Fig.1).1). Subsequently, the new methylene side chain would be epoxidated, followed by opening to an aldehyde and subsequent reduction to a hydroxyl (see steps 5 to 7 in Fig. Fig.1).1). The point of contention concerning this scheme is that it would also result in a 1,2-H shift between C-1 and C-5, which has not been observed (47). Nevertheless, the retention of only one methionine methyl-derived hydrogen, the 1,2-H shift between C-5 and C-6, and the lack of a 1,2-H shift between C-1 and C-5 are entirely consistent with our genetic revision of the STX-biosynthesis pathway, whereby the introduction of methionine methyl precedes the formation of the three heterocycles.

sxtD encodes an enzyme with sequence similarity to sterol desaturase and is the only candidate desaturase present in the sxt gene cluster. SxtD is predicted to introduce a double bond between C-1 and C-5 of C′, resulting in the 1,2-H shift between C-5 and C-6 (compound D′; Fig. Fig.3).3). The gene product of sxtS has sequence homology to nonheme iron 2-oxoglutarate-dependent dioxygenases. These are multifunctional enzymes that can perform hydroxylation, epoxidation, desaturation, cyclization, and expansion reactions (42). 2-Oxoglutarate dioxygenases have been reported to catalyze the oxidative formation of heterocycles (61). SxtS could therefore perform the consecutive epoxidation of the new double bond and opening of the epoxide to an aldehyde with concomitant bicyclization. This would explain the retention of only one methionine methyl-derived hydrogen and the lack of a 1,2-H shift between C-1 and C-5 of STX (see steps 5 to 7 in Fig. Fig.3).3). SxtU has sequence similarity to short-chain alcohol dehydrogenases. The most similar enzyme with a known function is clavaldehyde dehydrogenase (NCBI accession no. AAF86624), which reduces the terminal aldehyde of clavulanate-9-aldehyde to an alcohol (13). SxtU is therefore predicted to reduce the terminal aldehyde group of the STX precursor in step 8 (Fig. (Fig.3),3), forming compound E′.

The concerted action of SxtD, SxtS, and SxtU is therefore responsible for the hydroxylation and bicyclization of compound C′ to E′ (Fig. (Fig.3).3). In support for this proposed revision of STX biosynthesis descriptions, LC-MS-MS obtained using m/z values of 211 and 225 allowed the detection of compounds C′ and E′ from C. raciborskii T3 (Fig. (Fig.4).4). On the other hand, no evidence could be found by LC-MS-MS for intermediates B (m/z, 216), and C (m/z, 198), as previously predicted (47) (Fig. (Fig.1).1). MS-MS spectra showed the expected fragment ions (58) after the loss of ammonia and guanidine from C′, as well as the loss of water in the case of E′. Compounds E′ and C were not detected in extracts from C. raciborskii AWT205.

The detection of E′ indicated that the final reactions leading to the complete STX molecule are the O-carbamoylation of its free hydroxy group and a double hydroxylation of C-12 (Fig. (Fig.1).1). The actual sequence of these final reactions, however, remains uncertain. The gene product of sxtI is most similar to a predicted O-carbamoyltransferase from Trichodesmium erythraeum (accession no. ABG50968) and other predicted O-carbamoyltransferases from cyanobacteria. O-carbamoyltransferases invariably transfer a carbamoyl group from carbamoylphosphate to a free hydroxyl group. Our data indicate that SxtI may catalyze the transfer of a carbamoyl group from carbamoylphosphate to the free hydroxy group of E′ (Fig. (Fig.1).1). Homologues of sxtJ and sxtK with a known function were not found in the databases; however, it was noted that sxtJ and sxtK homologues were often encoded adjacent to O-carbamoyltransferase genes.

The sxt gene cluster contains two genes, sxtH and sxtT, each encoding a terminal oxygenase subunit of bacterial phenylpropionate and related ring-hydroxylating dioxygenases. The closest homologue with a predicted function was capreomycidine hydroxylase from Streptomyces vinaceus (VioQ; accession no. AAP92508) (54), which hydroxylates a ring carbon (C-6) of capreomycidine. SxtH and SxtT may therefore perform a similar function in STX biosynthesis, that is, the consecutive hydroxylation of C-12, converting F′ into STX (see step 10 in Fig. Fig.11).

Members belonging to bacterial phenylpropionate and related ring-hydroxylating dioxygenases are multicomponent enzymes, as they require an oxygenase reductase for their regeneration after each catalytic cycle. The sxt gene cluster provides a putative electron transport system, which would fulfill this function. sxtV encodes a 4Fe-4S ferredoxin with high sequence homology to a ferredoxin from Nostoc punctiforme (accession no. ZP_00106179). sxtW was most similar to fumarate reductase/succinate dehydrogenase-like enzymes from Anabaena variabilis (ABA24604) and Nostoc punctiforme (accession no. ZP_00106180), followed by AsfA from Pseudomonas putida (AAD31780). AsfA and AsfB are enzymes involved in the transport of electrons resulting from the catabolism of aryl sulfonates (57). SxtV could putatively extract an electron pair from succinate, converting it to fumarate (17), and then transfer the electrons via ferredoxin (SxtW) to SxtH and SxtT.

Comparative sequence analysis and functional assignment of tailoring genes.

Following synthesis of the parent molecule STX, modifying enzymes introduce various functional groups. In addition to STX, C. raciborskii T3 produces N-1-hydroxylated (neoSTX), decarbamoylated (dcSTX), and N-sulfurylated (GTX-5) toxins, whereas Anabaena circinalis AWQC131C produces decarbamoylated (dcSTX) toxins and O-sulfurylated (GTX-3/GTX-2, dcGTX-3/dcGTX-2) toxins, as well as both O- and N-sulfurylated toxins (C-1/C-2), but no N-1-hydroxylated toxins (25).

sxtX encodes an enzyme with homology to cephalosporin hydroxylase. sxtX was detected only in C. raciborskii T3, Aphanizomenon flos-aquae NH-5, and Lyngbya wollei, which produce N-1-hydroxylated analogues of STX (7, 21, 26), such as neoSTX. This component of the gene cluster was not present in any strain of A. circinalis, and therefore probably represents the reason why this species does not produce N-1-hydroxylated PSP toxins (25, 56). The predicted function of SxtX is therefore the N-1 hydroxylation of STX.

A. circinalis AWQC131C and C. raciborskii T3 also produce N- and O-sulfated analogues of STX [GTX-5, C-2/C-3, (dc)GTX-3/GTX-4]. The activity of two 3′-phosphate 5′-phosphosulfate (PAPS)-dependent sulfotransferases, which were specific for the N-21 of STX and GTX-3/GTX-2 and the O-22 of 11-hydroxy STX, respectively, has been described previously in studies of the PSP toxin-producing dinoflagellate Gymnodinium catenatum (44, 62). The sxt gene cluster from C. raciborskii T3 encodes SxtN, a putative sulfotransferase. A PSI-BLAST search using SxtN identified only 25 hypothetical proteins of unknown function with an E value above the threshold (0.005). A profile library search, however, revealed significant structural relatedness of SxtN to estrogen sulfotransferase (1AQU) (Z score = 24.02) and other sulfotransferases. SxtN has a conserved N-terminal region, which corresponds to the adenosine PAPS-binding region in 1AQU (20). It is not known, however, whether SxtN transfers a sulfate group to N-21 or O-22. Interestingly, the sxt gene cluster encodes an adenylylsulfate kinase (APSK), SxtO, homologues of which are involved in the formation of PAPS (Table (Table2).2). APKS phosphorylates adenylylsulfate, the product of ATP-sulfurylase, converting it to PAPS. Other biosynthetic gene clusters that result in sulfated secondary metabolites also contain genes required for the production of PAPS (45).

Decarbamoylated analogues of STX could be produced via either of two hypothetical scenarios. Enzymes that act downstream of SxtI, the carbamoyltransferase, in the biosynthesis of PSP toxins are proposed to exhibit broad substrate specificity, processing both carbamoylated and decarbamoylated precursors of STX. Alternatively, hydrolytic cleavage of the carbamoyl moiety from STX or its precursors may occur. SxtL is related to GDSL lipases, which are multifunctional enzymes with thioesterase, arylesterase, protease, and lysophospholipase activities (1). The function of SxtL could therefore include the hydrolytic cleavage of the carbamoyl group from STX analogues.

Cluster-associated genes involved in metabolite transport.

Kinetic studies of PSP toxin accumulation in producing cells and the media of cyanobacterial cultures suggest that there is an active transport mechanism for these toxins (9). In addition, variations in the concentration of sodium in culture media are known to affect the accumulation of PSP toxins in producer cells (41). sxtF and sxtM encoded two proteins with high sequence similarity to sodium-driven multidrug and toxic compound extrusion (MATE) proteins of the NorM family. Members of the NorM family of MATE proteins are bacterial sodium-driven antiporters that export cationic substances (6). All of the PSP toxins are cationic substances, except for the C toxins, which are zwitterionic. It is therefore probable that SxtF and SxtM are also involved in the export of PSP toxins. A mutational study of NorM from Vibrio parahaemolyticus identified three conserved negatively charged residues (D32, E251, and D367) that confer substrate specificity; however, the mechanism of substrate recognition remains unknown (34). In SxtF, the residue corresponding to E251 of NorM is conserved, whereas those corresponding to D32 and D367 are replaced by the neutral amino acids asparagine and tyrosine, respectively. Residues corresponding to D32 and E251 are conserved in SxtM, but D367 is replaced by histidine. The changes in substrate-binding residues may reflect the differences in PSP toxin substrates transported by these proteins.

Putative transcriptional regulators of saxitoxin synthase.

Environmental factors such as nitrogen and phosphate availability have been reported to regulate the production of PSP toxins in dinoflagellates and cyanobacteria (3, 10). Two transcriptional factors, sxtY and sxtZ, related to PhoU and OmpR, respectively, as well as a two-component regulator histidine kinase proximal to the 3′ end of the sxt gene cluster in C. raciborskii T3 have been identified. PhoU-related proteins are negative regulators of phosphate uptake (52), whereas OmpR-like proteins are involved in the regulation of a variety of metabolisms, including nitrogen (11) and osmotic balance (12). It is therefore likely that PSP toxin production in C. raciborskii T3 may be regulated at the transcriptional level in response to the availability of phosphate as well as other environmental factors.

Phylogenetic origins of the sxt genes.

There is considerable interest regarding the phylogenetic origin of PSP toxin biosynthesis and how the capacity to produce these toxins has developed in such distantly related organisms, namely, dinoflagellates and bacteria (38). The sxt gene cluster from C. raciborskii T3 has a true mosaic structure. Approximately half of the sxt genes of C. raciborskii T3 were most similar to counterparts from other cyanobacteria; however, the remaining genes had their closest matches with homologues from proteobacteria, actinomycetes, sphingobacteria, and firmicutes. There is an increasing body of evidence indicating that horizontal gene transfer (HGT) is a major driving force behind the evolution of prokaryotic genomes (14), and cyanobacterial genomes are known to be greatly affected by HGT, often involving transposases and phages (35). The fact that the majority of sxt genes are most closely related to homologues from other cyanobacteria suggests that STX biosynthesis may have evolved in an ancestral cyanobacterium that successively acquired the remaining genes from other bacteria via HGT. The structural organization of the investigated sxt gene cluster, as well as the presence of several transposases related to the IS4 family (8), suggests that small cassettes of sxt genes are mobile. The reasons for the restricted distribution of these biosynthetic genes to only four species in two orders of the Cyanobacteria are unknown but are critical for understanding the evolution of additional toxigenic species in the future. Sequencing of the sxt genes from several dinoflagellate species is under way and will assist in the further elucidation of the origin and evolution of these genes.

The constitutive expression of sxt genes in PSP toxin-producing cyanobacterial strains via reverse transcriptase PCR has been verified (data not shown), and the transcriptional regulation of the biosynthetic cluster is also under investigation. These experiments will aid the elucidation of the physiological function of the saxitoxins in bacteria and algae.

Saxitoxin synthase is a complex multienzyme pathway involving a novel PKS. The discovery of these genes will enable the rapid and accurate detection of harmful PSP-producing species in water and seafood. Intermediate compound analysis by MS and in vitro biosynthesis confirmed the sequence of reactions. Saxitoxin synthesis is the first nonterpene alkaloid pathway described for bacteria. Apart from understanding the functional role of these secondary metabolites and the factors that regulate their expression in the environment, accurate descriptions of the architecture of this complex system should provide strategies for combinatorial biosynthesis to generate novel neuroactive products.

Supplementary Material

[Supplemental material]


The Australian Research Council is thanked for its financial support of this work in the form of a postdoctoral fellowship to Y.J.J. and a professorial fellowship and discovery grant to B.A.N. A University of New South Wales Vice-Chancellor's Postdoctoral Fellowship to F.P. is also acknowledged. MS analyses were performed at the Bioanalytical Mass Spectrometry Facility (University of New South Wales), supported in part by grants from the Australian Government Systemic Infrastructure Initiative and Major National Research Facilities Program (UNSW node of the Australian Proteome Analysis Facility) and by the University of New South Wales Capital Grants Scheme and the Faculty of Engineering Research Grant for 2005.

Wayne Carmichael and Peter Baker are thanked for providing cyanobacterial samples.


Published ahead of print on 16 May 2008.

Supplemental material for this article may be found at http://aem.asm.org/.


1. Akoh, C. C., G. C. Lee, Y. C. Liaw, T. H. Huang, and J. F. Shaw. 2004. GDSL family of serine esterases/lipases. Prog. Lipid Res. 43:534-552. [PubMed]
2. Alexander, F. W., E. Sandmeier, P. K. Mehta, and P. Christen. 1994. Evolutionary relationships among pyridoxal-5′-phosphate-dependent enzymes. Regio-specific alpha, beta and gamma families. Eur. J. Biochem. 219:953-960. [PubMed]
3. Anderson, D. M., D. M. Kulis, J. J. Sullivan, S. Hall, and C. Lee. 1990. Dynamics and physiology of saxitoxin production by the dinoflagellates Alexandrium spp. Mar. Biol. 104:511-524.
4. Arakawa, O., S. Nishio, T. Noguchi, Y. Shida, and Y. Onoue. 1995. A new saxitoxin analogue from a xanthid crab Atergatis floridus. Toxicon 33:1577-1584. [PubMed]
5. Arakawa, O., T. Noguchi, Y. Shida, and Y. Onoue. 1994. Occurrence of carbamoyl-N-hydroxy derivatives of saxitoxin and neosaxitoxin in a xanthid crab Zosimus aeneus. Toxicon 32:175-183. [PubMed]
6. Brown, M. H., I. T. Paulsen, and R. A. Skurray. 1999. The multidrug efflux protein NorM is a prototype of a new family of transporters. Mol. Microbiol. 31:394-395. [PubMed]
7. Carmichael, W. W., W. R. Evans, Q. Q. Yin, P. Bell, and E. Moczydlowski. 1997. Evidence for paralytic shellfish poisons in the freshwater cyanobacterium Lyngbya wollei (Farlow ex Gomont) comb. nov. Appl. Environ. Microbiol. 63:3104-3110. [PMC free article] [PubMed]
8. Cassier-Chauvat, C., M. Poncelet, and F. Chauvat. 1997. Three insertion sequences from the cyanobacterium Synechocystis PCC6803 support the occurrence of horizontal DNA transfer among bacteria. Gene 195:257-266. [PubMed]
9. Castro, D., D. Vera, N. Lagos, C. Garcia, and M. Vasquez. 2004. The effect of temperature on growth and production of paralytic shellfish poisoning toxins by the cyanobacterium Cylindrospermopsis raciborskii C10. Toxicon 44:483-489. [PubMed]
10. Dias, E., P. Pereira, and S. Franca. 2002. Production of paralytic shellfish toxins by Aphanizomenon sp. LMECYA 31 (Cyanobacteria). J. Phycol. 38:705-712.
11. Ehira, S., and M. Ohmori. 2006. NrrA, a nitrogen-responsive response regulator, facilitates heterocyst development in the cyanobacterium Anabaena sp. strain PCC 7120. Mol. Microbiol. 59:1692-1703. [PubMed]
12. Forst, S., J. Delgado, and M. Inouye. 1989. Phosphorylation of OmpR by the osmosensor EnvZ modulates expression of the ompF and ompC genes in Escherichia coli. Proc. Natl. Acad. Sci. USA 86:6052-6056. [PMC free article] [PubMed]
13. Fulston, M., M. Davison, S. W. Elson, N. H. Nicholson, J. W. Tyler, and S. R. Woroniecki. 2001. Clavulanic acid biosynthesis; the final steps. J. Chem. Soc. Perkin Trans. 1:1122-1130.
14. Gogarten, J. P., and J. P. Townsend. 2005. Horizontal gene transfer, genome innovation and evolution. Nat. Rev. Microbiol. 3:679-687. [PubMed]
15. Hackett, J. D., T. E. Scheetz, H. S. Yoon, M. B. Soares, M. F. Bonaldo, T. L. Casavant, and D. Bhattacharya. 2005. Insights into a dinoflagellate genome through expressed sequence tag analysis. BMC Genomics 6:80. [PMC free article] [PubMed]
16. Hallegraeff, G. M. 1995. Harmful algal blooms: a global overview, p. 1-22. In G. M. Hallegraeff, D. M. Anderson, and A. D. Cembella (ed.), Manual on harmful marine microalgae. UNESCO, Paris, France. (In French.)
17. Hederstedt, L., and L. Rutberg. 1981. Succinate dehydrogenase—a comparative review. Microbiol. Rev. 45:542-555. [PMC free article] [PubMed]
18. Kaas, H., and P. Henriksen. 2000. Saxitoxins (PSP toxins) in Danish lakes. Water Res. 34:2089-2097.
19. Kagan, R. M., and S. Clarke. 1994. Widespread occurrence of three sequence motifs in diverse S-adenosylmethionine-dependent methyltransferases suggests a common structure for these enzymes. Arch. Biochem. Biophys. 310:417-427. [PubMed]
20. Kakuta, Y., L. G. Pedersen, C. W. Carter, M. Negishi, and L. C. Pedersen. 1997. Crystal structure of estrogen sulphotransferase. Nat. Struct. Biol. 4:904-908. [PubMed]
21. Kao, C. Y., and S. R. Levinson (ed.). 1986. Tetrodotoxin, saxitoxin, and the molecular biology of the sodium channel. Ann. N. Y. Acad. Sci. 479:1-445. [PubMed]
22. Kellmann, R., and B. A. Neilan. 2007. Biochemical characterization of paralytic shellfish toxin biosynthesis in vitro. J. Phycol. 43:497-508.
23. Lagos, N., H. Onodera, P. A. Zagatto, D. Andrinolo, S. Azevedo, and Y. Oshima. 1999. The first evidence of paralytic shellfish toxins in the freshwater cyanobacterium Cylindrospermopsis raciborskii, isolated from Brazil. Toxicon 37:1359-1373. [PubMed]
24. Lambalot, R. H., A. M. Gehring, R. S. Flugel, P. Zuber, M. Lacelle, M. A. Marahiel, R. Reid, C. Khosla, and C. T. Walsh. 1996. A new enzyme superfamily—the phosphopantetheinyl transferases. Chem. Biol. 3:923-936. [PubMed]
25. Llewellyn, L. E., A. P. Negri, J. Doyle, P. D. Baker, E. C. Beltran, and B. A. Neilan. 2001. Radioreceptor assays for sensitive detection and quantitation of saxitoxin and its analogues from strains of the freshwater cyanobacterium, Anabaena circinalis. Environ. Sci. Technol. 35:1445-1451. [PubMed]
26. Mahmood, N. A., and W. W. Carmichael. 1986. Paralytic shellfish poisons produced by the freshwater cyanobacterium Aphanizomenon flos-aquae NH-5. Toxicon 24:175-186. [PubMed]
27. Moffitt, M. C., and B. A. Neilan. 2004. Characterization of the nodularin synthetase gene cluster and proposed theory of the evolution of cyanobacterial hepatotoxins. Appl. Environ. Microbiol. 70:6353-6362. [PMC free article] [PubMed]
28. Negri, A., D. Stirling, M. Quilliam, S. Blackburn, C. Bolch, I. Burton, G. Eaglesham, K. Thomas, J. Walter, and R. Willis. 2003. Three novel hydroxybenzoate saxitoxin analogues isolated from the dinoflagellate Gymnodinium catenatum. Chem. Res. Toxicol. 16:1029-1033. [PubMed]
29. Neilan, B. A. 1995. Identification and phylogenetic analysis of toxigenic cyanobacteria by multiplex randomly amplified polymorphic DNA PCR. Appl. Environ. Microbiol. 61:2286-2291. [PMC free article] [PubMed]
30. Neilan, B. A., M. L. Saker, J. Fastner, A. Torokne, and B. P. Burns. 2003. Phylogeography of the invasive cyanobacterium Cylindrospermopsis raciborskii. Mol. Ecol. 12:133-140. [PubMed]
31. Neuwald, A. F., and D. Landsman. 1997. GCN5-related histone N-acetyltransferases belong to a diverse superfamily that includes the yeast SPT10 protein. Trends Biochem. Sci. 22:154-155. [PubMed]
32. Onodera, H., M. Satake, Y. Oshima, T. Yasumoto, and W. Carmichael Wayne. 1997. New saxitoxin analogues from the freshwater filamentous cyanobacterium Lyngbya wollei. Nat. Toxins 5:146-151. [PubMed]
33. Oshima, Y. 1995. Postcolumn derivatization liquid chromatographic method for paralytic shellfish toxins. J. AOAC Int. 78:528-532.
34. Otsuka, M., M. Yasuda, Y. Morita, C. Otsuka, T. Tsuchiya, H. Omote, and Y. Moriyama. 2005. Identification of essential amino acid residues of the NorM Na+/multidrug antiporter in Vibrio parahaemolyticus. J. Bacteriol. 187:1552-1558. [PMC free article] [PubMed]
35. Palenik, B., B. Brahamsha, F. W. Larimer, M. Land, L. Hauser, P. Chain, J. Lamerdin, W. Regala, E. E. Allen, J. McCarren, I. Paulsen, A. Dufresne, F. Partensky, E. A. Webb, and J. Waterbury. 2003. The genome of a motile marine Synechococcus. Nature 424:1037-1042. [PubMed]
36. Pereira, P., H. Onodera, D. Andrinolo, S. Franca, F. Araujo, N. Lagos, and Y. Oshima. 2000. Paralytic shellfish toxins in the freshwater cyanobacterium Aphanizomenon flos-aquae, isolated from Montargil reservoir, Portugal. Toxicon 38:1689-1702. [PubMed]
37. Piel, J., G. Wen, M. Platzer, and D. Hui. 2004. Unprecedented diversity of catalytic domains in the first four modules of the putative pederin polyketide synthase. Chembiochem 5:93-98. [PubMed]
38. Plumley, F. G. 2001. Purification of an enzyme involved in saxitoxin synthesis. J. Phycol. 37:926-928.
39. Pomati, F., B. P. Burns, and B. A. Neilan. 2004. Identification of an Na+-dependent transporter associated with saxitoxin-producing strains of the cyanobacterium Anabaena circinalis. Appl. Environ. Microbiol. 70:4711-4719. [PMC free article] [PubMed]
40. Pomati, F., and B. A. Neilan. 2004. PCR-based positive hybridization to detect genomic diversity associated with bacterial secondary metabolism. Nucleic Acids Res. 32:e7. [PMC free article] [PubMed]
41. Pomati, F., C. Rossetti, G. Manarolla, B. P. Burns, and B. A. Neilan. 2004. Interactions between intracellular Na+ levels and saxitoxin production in Cylindrospermopsis raciborskii T3. Microbiology 150:455-461. [PubMed]
42. Prescott, A. G., and M. Lloyd. 2000. The iron(II) and 2-oxoacid-dependent dioxygenases and their role in metabolism. Nat. Prod. Rep. 17:367-383. [PubMed]
43. Saker, M. L., and D. J. Griffiths. 2000. The effect of temperature on growth and cylindrospermopsin content of seven isolates of Cylindrospermopsis raciborskii (Nostocales, Cyanophyceae) from water bodies in northern Australia. Phycologia 39:349-354.
44. Sako, Y., T. Yoshida, A. Uchida, O. Arakawa, T. Noguchi, and Y. Ishida. 2001. Purification and characterization of a sulfotransferase specific to N-21 of saxitoxin and gonyautoxin 2+3 from the toxic dinoflagellate Gymnodinium catenatum (Dinophyceae). J. Phycol. 37:1044-1051.
45. Schwedock, J. S., and S. R. Long. 1992. Rhizobium meliloti genes involved in sulfate activation: the two copies of nodPQ and a new locus, saa. Genetics 132:899-909. [PMC free article] [PubMed]
46. Shaw, G. R., A. Sukenik, A. Livne, R. K. Chiswell, M. J. Smith, A. A. Seawright, R. L. Norris, G. K. Eaglesham, and M. R. Moore. 1999. Blooms of the cylindrospermopsin containing cyanobacterium, Aphanizomenon ovalisporum (Forti), in newly constructed lakes, Queensland, Australia. Environ. Toxicol. 14:167-177.
47. Shimizu, Y. 1993. Microalgal metabolites. Chem. Rev. 93:1685-1698.
48. Shimizu, Y., M. Norte, A. Hori, A. Genenah, and M. Kobayashi. 1984. Biosynthesis of saxitoxin analogues: the unexpected pathway. J. Am. Chem. Soc. 106:6433-6434.
49. Siebert, P. D., A. Chenchik, D. E. Kellogg, K. A. Lukyanov, and S. A. Lukyanov. 1995. An improved PCR method for walking in uncloned genomic DNA. Nucleic Acids Res. 23:1087-1088. [PMC free article] [PubMed]
50. Sleno, L., D. Volmer, and A. Marshall. 2005. Assigning product ions from complex MS/MS spectra: the importance of mass uncertainty and resolving power. J. Am. Soc. Mass Spectrom. 16:183-198. [PubMed]
51. Sleno, L., D. A. Volmer, B. Kovacevic, and Z. B. Maksic. 2004. Gas-phase dissociation reactions of protonated saxitoxin and neosaxitoxin. J. Am. Soc. Mass Spectrom. 15:462-477. [PubMed]
52. Steed, P. M., and B. L. Wanner. 1993. Use of the rep technique for allele replacement to construct mutants with deletions of the pstSCAB-phoU operon: evidence of a new role for the PhoU protein in the phosphate regulon. J. Bacteriol. 175:6797-6809. [PMC free article] [PubMed]
53. Su, Z., M. Sheets, H. Ishida, F. Li, and W. H. Barry. 2004. Saxitoxin blocks L-type ICa. J. Pharmacol. Exp. Ther. 308:324-329. [PubMed]
54. Thomas, M. G., Y. A. Chan, and S. G. Ozanick. 2003. Deciphering tuberactinomycin biosynthesis: isolation, sequencing, and annotation of the viomycin biosynthetic gene cluster. Antimicrob. Agents Chemother. 47:2823-2830. [PMC free article] [PubMed]
55. Thompson, A. S., J. C. Rhodes, and I. Pettman. 1988. Culture collection of algae and protozoa: catalogue of strains, 5th ed, p. 22. Natural Environment Research Council, Freshwater Biological Association, Ambleside, United Kingdom.
56. Velzeboer, R. M. A., P. D. Baker, J. Rositano, T. Heresztyn, G. A. Codd, and S. L. Raggett. 2000. Geographical patterns of occurrence and composition of saxitoxins in the cyanobacterial genus Anabaena (Nostocales, Cyanophyta) in Australia. Phycologia 39:395-407.
57. Vermeij, P., C. Wietek, A. Kahnert, T. Wueest, and M. A. Kertesz. 1999. Genetic organization of sulfur-controlled aryl desulfonation in Pseudomonas putida S-313. Mol. Microbiol. 32:913-916. [PubMed]
58. Vishwanathan, K., R. L. Tackett, J. T. Stewart, and M. G. Bartlett. 2000. Determination of arginine and methylated arginines in human plasma by liquid chromatography-tandem mass spectrometry. J. Chromatogr. B Biomed. Sci. Appl. 748:157-166. [PubMed]
59. Wang, J., J. J. Salata, and P. B. Bennett. 2003. Saxitoxin is a gating modifier of HERG K+ channels. J. Gen. Physiol. 121:583-598. [PMC free article] [PubMed]
60. Wilson, K. M., M. A. Schembri, P. D. Baker, and C. P. Saint. 2000. Molecular characterization of the toxic cyanobacterium Cylindrospermopsis raciborskii and design of a species-specific PCR. Appl. Envir. Microbiol. 66:332-338. [PMC free article] [PubMed]
61. Yin, X., T. O'Hare, S. J. Gould, and T. M. Zabriskie. 2003. Identification and cloning of genes encoding viomycin biosynthesis from Streptomyces vinaceus and evidence for involvement of a rare oxygenase. Gene 312:215-224. [PubMed]
62. Yoshida, T., Y. Sako, A. Uchida, T. Kakutani, O. Arakawa, T. Noguchi, and Y. Ishida. 2002. Purification and characterization of sulfotransferase specific to O-22 of 11-hydroxy saxitoxin from the toxic dinoflagellate Gymnodinium catenatum (Dinophyceae). Fish. Sci. 68:634-642.
63. Zaman, L., O. Arakawa, A. Shimosu, Y. Shida, and Y. Onoue. 1998. Occurrence of a methyl derivative of saxitoxin in Bangladeshi freshwater puffers. Toxicon 36:627-630. [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem chemical compound records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records. Multiple substance records may contribute to the PubChem compound record.
  • MedGen
    Related information in MedGen
  • Nucleotide
    Primary database (GenBank) nucleotide records reported in the current articles as well as Reference Sequences (RefSeqs) that include the articles as references.
  • PopSet
    Sets of sequences from population and evolutionary genetic studies in the PopSet database reported in the current articles.
  • Protein
    Protein translation features of primary database (GenBank) nucleotide records reported in the current articles as well as Reference Sequences (RefSeqs) that include the articles as references.
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem chemical substance records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records.

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...