The toxigenic freshwater cyanobacterium Cylindrospermopsis raciborskii T3 has been used as a model to study and elucidate the biosynthetic pathway of tetrahydropurine neurotoxins associated with paralytic shellfish poisoning (PSP). There are nevertheless several inconsistencies and contradictions in the toxin profile of this strain as published by different research groups, and claimed to include carbamoyl (SIX, NEO, GTX2/3), decarbamoyl (dcSTX), and N-sulfocarbamoyl (C1/2, 81) derivatives. Our analysis of the complete genome of another PSP toxin-producing cyanobacterium, Raphidiopsis brookii D9, which is closely related to C. raciborskii T3, resolved many issues regarding the correlation between biosynthetic pathways, corresponding genes and the T3 toxin profile. The putative sxt gene cluster in R. brookii D9 has a high synteny with the T3 sxt cluster, with 100% nucleotide identity among the shared genes. We also compared the PSP toxin profile of the strains by liquid chromatography coupled to mass spectrometry (LC-MS/MS). In contrast to published reports, our reassessment of the PSP toxin profile of T3 confirmed production of only SIX, NEO and dcNEO. We gained significant insights via correlation between specific sxt genes and their role in PSP toxin synthesis in both D9 and T3 strains. In particular, analysis of sulfotransferase functions for SxtN (N-sulfotransferase) and SxtSUL (O-sulfotransferase) enzymes allowed us to propose an extension of the PSP toxin biosynthetic pathway from SIX to the production of the derivatives GTX2/3. C1/2 and B1. This is a significantly revised view of the genetic mechanisms underlying synthesis of sulfated and sulfonated STX analogues in toxigenic cyanobacteria. (C) 2010 Elsevier Ltd. All rights reserved.