Conceived and designed the experiments: RFS PJK. Performed the experiments: RFS MAC ERJ. Analyzed the data: RFS AH RA MAC FCK IJ MH ERJ PJK. Contributed reagents/materials/analysis tools: AH RA MAC RAA FCK IJ MH BV FK JB PJK. Wrote the paper: RFS PJK. Contributed to manuscript preparation: AH RA MAC FCK RAA IJ MH BV FK JB. Established cultures: MAC MH. Contributed to collective acquisition of cultures: RAA. Contributed to interpretation of results: RAA IJ MH. Contributed to preparation of materials: IJ.
The authors have declared that no competing interests exist.
Dinoflagellates are an ecologically important group of protists with important functions as primary producers, coral symbionts and in toxic red tides. Although widely studied, the natural diversity of dinoflagellates is not well known. DNA barcoding has been utilized successfully for many protist groups. We used this approach to systematically sample known “species”, as a reference to measure the natural diversity in three marine environments.
In this study, we assembled a large cytochrome
COI barcoding was successful in identifying species from 70% of cultured genera. When applied to environmental samples, it revealed a massive amount of natural diversity in dinoflagellates. This highlights the extent to which we underestimate microbial diversity in the environment.
Assessing biodiversity in the microbial world has always been a difficult problem: not only are microorganisms inherently more difficult to examine and differentiate by classical methods, but it is also not clear if the theoretical taxonomic frameworks, applied to more familiar life forms, even apply to the diversity of microbial life. Even the validity of the species concept is debatable for some protist groups. Within microbial eukaryotes, the protists, there is a persistent debate over how much diversity exists, irrespective of how we divide it up. On one side of the debate it is argued that protist diversity typically consists of a relatively few cosmopolitan species because their small size allows them to live ubiquitously
The application of molecular systematics to various protist lineages has revealed unexpected levels of diversity and a surge in the documentation of morphologically cryptic species
Dinoflagellates are an ancient and evolutionarily complex group of protists, members of which occupy every major ecological niche from primary producers to parasites (reviewed in
Recently, DNA barcoding was used to assess freshwater and brackish dinoflagellates using two mitochondrial markers, a small number of COI (cytochrome
Our survey successfully identified 101 strains (cultures with separate identification labels) from species belonging to 15 of 18 genera from culture collections to species level with a good correlation between named species and its COI sequence in most sub-groups. Nevertheless, several cases of cryptic diversity within culture collections were evident, particularly in the genus
Out of 669 culture collection samples from 11 collections, we retrieved 566 COI amplicons as some taxa failed to amplify (most commonly, these were
Average pairwise distances (PWD) were calculated for all strains with named species within 18 genera to measure variance within species over the whole dataset in order to account for differing sample sizes, ranging from 1 to 16 strains per species (average 3.4) and 6
Horizontal axis compares strain isolates within a species (or strains within a clade for
Not surprisingly, the genus
Comparing species within a genus revealed large variation from zero to just over 8% (see
Many strains in culture collections have not been identified to the level of species, and our analyses have successfully identified 101 of these to known species. Excluding strains that were refractory to COI barcoding due to poor resolution (see below), differences between barcodes and species or genus identity were encountered in 17 strains (see
Within the recently identified Pfiesteria group that inhabits brackish water
Each species is coloured according to its original species or clade designation. Species of the same genus or clades share a colour theme. Brackets indicate “barcode” species groupings, calculated by uncorrected pairwise distance of 0.24% or less, including those for
Cryptic diversity is also evident within
In other cases, taxonomically distinct species were found to have identical DNA barcodes. In our dataset, eight species showed no distinction by COI barcode.
In contrast, other taxonomically related species were found to be quite distinct. The two
Takabayashi and colleagues
Brackets indicate all strains within a minimal barcode genus-O.T.U, with a PWD cut off of 1.4%. Pink shaded area indicates strains belonging to the Kareniaceae family. Where environmental barcodes (EB), could not be identified to any known strain in our database, they were labeled with a numbered EB genus, except for strains belonging to Kareniaceae family where they were prefixed Kar-fam., and then an assigned genus-O.T.U number or the name of most related known strain. All strains are listed in
Group A3 included two sequences belonging to clade A3, but also contained two strains that were identified to Cp23S-rDNA genotype marker A194 (nomenclature based on clade A and size of the hypervariable region, 194 bp) that generally corresponds to A1 but, may also belong to clade A4. Group Ax contained a mixture of strains belonging to A194 and a second, uncharacterized Cp23S-rDNA clade A genotype called A188 and contained all full-length COI clade A sequences deposited in Genbank. Strain Zs (AY289692) was identical to CCMP2461 (GQ501337) and had borderline identity (0.24–0.27%) to CCMP2429 (GQ501395), a more distal clade A2 strain (
Clade C/F contained two strains, CCMP2466 (GQ501334 clade C1) and Mv (AY289712) originally identified as F1
Pseudogenes are known to occur in dinoflagellates
In order to evaluate how much natural diversity is represented in culture collections, environmental DNA barcoding of total planktonic DNA from three different marine environments was carried out with dinoflagellate-specific primer sets. The deepest sampling was done from Saanich Inlet - a marine fjord off Vancouver Island that is hypoxic from 100 m to 200 m depth in the summer, but which is mixed in the winter. Near-coastal planktonic samples were also taken from the coast of Maine, USA and the island of Guadeloupe in the Caribbean. After screening out poor quality sequences, this resulted in a total of 713 environmental barcode sequences (listed in
Clustering the combined culture collection and environmental data resulted in a total of 1049 dinoflagellate barcodes (
Overall, the majority of environmental barcodes from these marine samples could not be identified because there was little overlap in diversity between the environmental and cultured dinoflagellate barcodes. In addition, there was little overlap of species assemblages between different environments. Nevertheless numerous other environmental barcodes could be identified to the genus level or higher, and many of these represent great expansions in the known diversity of these groups.
With the majority of environmental barcodes not attributable to any taxonomic group represented in the barcode library of culture collections, a method to connect environmental barcodes to cells will be required to assess natural microbial diversity. For a start, barcoding from single cells that have been photographed prior to isolation would at least allow us to identify major groups otherwise only made up of environmental sequences. To test this in principle, almost 70 single cells were photographed and manually isolated from the west coast of Vancouver Island, and 24 COI barcodes were generated from these single cells, although images for four single-cells in this study (GQ502036, HM236194, GQ502039, HM236195, GQ501405) were very poor quality and discarded. The single-cell barcodes proved to be successful in providing benchmarks for the environmental barcode libraries for the difficult-to-culture heterotrophic
Black solid lines show species level similarity.
Only four species of Kareniaceae are represented in our barcodes from culture collections (
The Kareniaceae have traditionally been a relatively small family made up of dinoflagellates that are distinguished by having a 19′ hexanoyloxyfucoxanthin-type plastid derived from a haptophyte
Culture collections represent the most accessible and traceable repositories of living microalgae, but they are also biased towards the species and strains most amenable to cultivation, and/or of commercial and medical importance. It has long been known, especially in prokaryotes, that only a small fraction of natural diversity is easily cultured using common strategies. Accordingly, cultured strains do not adequately represent either the depth or breadth of microbial diversity, but they still make useful benchmarks for molecular barcode databases of natural diversity. However, while it is becoming increasingly easy to use molecular data for taxonomic surveys that avoid culture work, cultures still remain a primary source for scientists who need cells for biochemical, biotechnical, cellular and physiological studies. These scientists will benefit greatly from barcodes that can provide an easy and quick means for quality control.
COI barcoding worked as a means of distinguishing most dinoflagellate species in 15 of 21 genera where COI sequence divergence rates were congruent with speciation events. For those 15 genera 81% of strains could be distinguished at species-O.T.U level using a PWD cut-off value of 0.24%. This criterion would allow for the majority of COI samples obtained from environmental samples to be binned into genetic clusters which on average corresponds with distinct species. This approach will overestimate the number of true species in genera with higher divergence rates and underestimate those with lower rates of divergence, but on average it should give a reasonable estimate of species diversity. A small assessment of species in the most variable genera demonstrated that genetic differences are unlikely to be due to pseudogenes or paralogues in all but one strain (discussed below), at least for those species tested. Other species, such as those belonging to Kareniaceae, showed low intra-species PWDs and unlikely to possess paralogues that would artificially inflate observed diversity in the environmental barcode dataset.
Assessing species diversity in the five genera (
It is difficult to compare the effectiveness of COI in comparison to cob
The genus
Whilst monoclonal cultured cells can be a useful source to control morphological plasticity
In common with previous SSU deep amplicon sequencing surveys of protists in marine environments
Recent studies described five new species of Kareniaceae
The summer and winter species compositions in Saanich Inlet were distinct and dominated by one genus. Coincidentally the greatest different in abiotic measurements were found between the February and July time points
Our results show cultured dinoflagellates that are considered to be different species can be resolved using DNA barcoding, although robust taxonomy using other DNA markers, morphology and chemotaxonomic markers such as lipids and pigments is needed to provide a solid basis for DNA barcoding to work
The environmental barcodes in this dataset may be a small fraction of the real diversity of the environments they represent, since none seemed to be sampled exhaustively. An even tinier proportion of dinoflagellates are represented by the combined holdings of culture collections, reflecting human bias in sampling and cultivation. DNA barcoding studies will be instrumental in evaluating biogeographical speciation and ecological assemblages within protist populations. With increasing use of next generation sequencing technology that can combine multiple markers, deep-level biodiversity studies will be more able to demonstrate true estimates of protist diversity and may provide useful information for the cultivation of a greater proportion of presently unculturable species. These methods combined with flow cytometry, already in use [
Cultured strains or DNA samples were donated or purchased from eight public and three private culture collections, summarized in
For environmental analysis, planktonic samples were collected through a 20 m plankton net taken from the island of Guadeloupe in the Caribbean (15.1539N, 61.3475W), from the Bigelow Laboratory pier, West Boothbay Harbor, ME, USA (38.1904N, 76.2707W) and also from mouth of the Damarascotta River, Maine, USA approximately 44N, 69.5W) representing two different coastal environments in Maine, USA (Northwest Atlantic). DNA from Saanich Inlet, Vancouver Island, BC, Canada (43.39N, 123.39W) (Northeast Pacific) was kindly donated by Dr. D. Walsh and Dr. S. Hallam, UBC, from non-filtered marine water collected by David Walsh in water collector containers at depths of 10 m, 100 m, 120 m and the bottom at 200 m in equal volumes. These same samples are also described by
Typically between 1.5–15 ml of dinoflagellate cells from culture were collected by centrifugation at initially 3000 g then at 1150 g, snap frozen in liquid nitrogen and thawed three times. For one third of culture collection samples, additional grinding was performed using plastic pestle and microfuge tube. DNA extraction was carried out using the DNeasy™ plant purification DNA kit (Qiagen, Mississauga, ON, Canada), following their protocol except incubating cells in lysis solution for 30 minutes instead of 10 minutes. Masterpure™ Complete DNA and RNA Purification Kit (Epicentre® Biotechnologies, Madison, WI, USA) was also used in about one third of cultures and for single cells, using Lysis of Fluid sample protocol followed by Precipitation of Total DNA protocol. For whole marine extracts DNA extraction was performed by mixing whole marine sample with an equal volume of a phenol-chloroform- isoamyl alcohol mixture (25∶24∶1) (Sigma-Aldrich, Oakville, ON, Canada), the DNA containing phase was removed and DNA extracted with the addition of 2.5 volumes of 100% ethanol (Sigma- Aldrich, Oakville, ON, Canada) and 0.1 volumes of 3M sodium acetate, pH 5.2 (Sigma-Aldrich), washed twice in 75% Ethanol and resuspended to 300 µg/µl in sterile water.
Highest amplification rates were achieved using a nested primer set, consisting of primers DINOCOX1F
COI barcodes for cultivated and uncultivated environmental dinoflagellates used in this study are listed in
In calculating genus-level O.T.U for environmental barcodes, we categorized each PWD comparison between species and ordered them according to increasing PWD. Each strain was checked and categorized so that every member of a genus-O.T.U group had no more than 1.4% identity to any other member of the same group. Although at the lower end of PWD of known genera, this value also minimized ambiguous identities, where a strain showed equal identity to more than one genus, or showed identity to only some but not all members of a genus-level O.T.U group. In these cases, the corresponding sequences were checked for sequence quality and excluded from analysis if below quality of other sequences. Even with sequence quality parsing, unresolved cases remained. These sequences were removed from their original group and placed in a separate smaller grouping or on their own, for single strains, as shown in
For
Neighbor Joining Cluster Analysis of Uncorrected PWD from All Culture Collection COI Barcodes as in
(1.58 MB TIF)
Neighbor Joining Cluster Analysis of Uncorrected PWD from Culture Collections and Marine Environmental COI Barcodes as in
(0.75 MB TIF)
Principle Coordinate separation of DNA Barcodes from Cultured and Environmental Dinoflagellates. PCoA using the first three principle coordinates (labeled all-st-1, 2 or 3) of all environmental barcodes and culture collection strains, colour-coded in a similar manner to
(0.75 MB TIF)
Identification of Cultivated and Genbank deposited Dinoflagellate Strains in this Study Using COI barcodes. PWD cut-off of Species-O.T.U is 0.24% or less, Genera 0.T.U is 1.4% or less. Strain Synonyms are indicated in brackets. Identification of COI barcode are explained in the
(0.16 MB XLS)
Uncultivated Environmental Barcodes in this Study from Dinoflagellate-specific Amplified DNA and from Individual Dinoflagellate Cells. O.T.Us were defined according to PWD comparison cut off values, based on values from known species comparisons (Species is 0.24% or less, genera is 1.4% or less). Naming of COI barcodes are explained in
(0.17 MB XLS)
Within group average calculations for seven species to determine presence of paralogues in COI barcodes. Five clones were sequenced per strain, except for CCMP1746 (3 clones) showing that clonal variation in all but one strain was less than 0.2%. Clones of CCMP421 revealed almost 10 times as much diversity compared to that of other dinoflagellates, including another
(0.03 MB XLS)
We thank all private and public culture collection for donating and participating in this study especially Donna Dinh for kindly donating samples from CCCM, Rebecca Gast and Dawn Moran for donating Antarctic dinoflagellate samples, Dion Frampton from ANACC as well as Cecilia Rad-Menendez, Christine N. Campbell and Joanne Field from CCAP for culture preparation and DNA preparation. Also thanks go to Julie Sexton from CCMP, Bertrand Le Roy from Algobank-Caen for culture preparation and assistance. We are grateful to Dr. S. Hallam and Dr. D. Walsh for donating DNA samples from Saanich Inlet. We thank B. Imanian for primer design and R.B. Moore for sharing unpublished results, Todd Harper for comments and review of this manuscript and Daniel Poland and Jill Mansfield from Professor Coffroth lab, State University of New York at Buffalo, USA for experimental analysis. We also wish to thank the Canadian Barcode of Life Network, Chris Grainger, Megan Milton, Natalia Ivanova and Sujeevan Ratnasingham for sequencing database and web support. We also thank Paul D. N. Hebert, Robert Hanner and Mehrdad Hajibabaei for their organizational support. Finally we extend our gratitude to thoughtful suggestions by the reviewers of this manuscript.