Phylogenetic analysis guided by intragenomic SSU rDNA polymorphism refines classification of Alexandrium tamarense species complex

We analyzed small subunit ribosomal rRNA gene (SSU rDNA) in single-cell isolated strains of Alexandnum tamarense, Alexandrium fundyense and Alexandrium catenella (Atama complex) by both direct and clone-based sequencing, and found 42-50 intragenomic SSU rDNA polymorphic sites (High IRP) in some of the strains but none or one (No/Low IRP) in others. Clone sequencing of the High-IRP amplicons revealed numerous variants with 0-3.5% nucleotide differences. Phylogenetic analyses including reported and new Atama complex SSU rDNA data divided this complex into two major well-separated clades. All the High-IRP sequences obtained in this study were grouped in Clade I ( High-IRP clade), which also contained strains of Atama complex previously reported from northern Asia. Clade II (No/Low-IRP clade), with no evidence of IRP except for the one polymorphic site found in one of the strains, contained a subclade (IIC) exclusively of A. catenella from various geographic locations and several other subclades (IIA, IIB) predominantly of A. tamarense. Clade I corresponded to Group I in the large subunit (LSU) rDNA-based phylogenetic tree (Lilly et al., 2007), and subclades in our Clade II corresponded to the LSU-based Groups II-V. Our IRP information further unites the many seemingly different genotypes into a coherent group (Clade I) and provides delineating boundary between this and other genotypes (Clade II) in Atama complex. Based on the currently available data, we propose that Clade I and subclade IIC represent two distinct species, while the rest of Clade II represents another one or more species. Clade I should be considered one species because (1) intragenomic rDNA variants dispersed across strains in the phylogenetic tree uniting the many different genotypes to same or closely related populations, (2) their expressed SSU rDNAs (i.e. rRNAs) are almost identical (1 out of 1700 nt, 0.06% difference), and (3) the corresponding Group I in the LSU-based tree also appears to be a coherent group. Subclade IIC should be considered another species because it has a long distance from the rest of Clade II as well as Clade I. Whether the rest of Clade II (mainly European A tamarense subclades) represents another one or more species requires further study. Our results show that the three original morphospecies designations are invalid, and the strains do not group based on geographic locations or whether they are toxic in general, although some subclades are predominated by part of a morphotype from a region. Applying the IRP-guided analysis to a toxic Alexandrium bloom in Long Island Sound revealed that although some undocumented Alexandrium-related and other dinoflagellate lineages coexisted, the majority of the diverse SSU sequences detected belonged to one single population identical to A. fundyense in Gulf of Maine. Our result suggests that negligence of IRP could lead to incorrect recognition of the intragenomic SSU rDNA variants as distinct genotypes, thus overestimating strain diversity of a bloom of this species complex and possibly other HAB lineages. (C) 2012 Elsevier B.V. All rights reserved.