J. Phycol. 47, 393–406 (2011) 2011 Phycological Society of America DOI: 10.1111/j.1529-8817.2011.00964.x

MOLECULAR PHYLOGENY OF DINOPHYSOID : THE SYSTEMATIC POSITION OF OXYTOXOIDES AND THE HASTATA GROUP (DINOPHYSALES, )1

Fernando Go´mez2 Universite´ Lille Nord de France, Laboratoire d¢Oce´anologie et Ge´osciences, CNRS UMR 8187, Station Marine de Wimereux, 28 Av. Foch, 62930 Wimereux, France Purificacio´nLo´pez-Garcı´a and David Moreira Unite´ d’Ecologie, Syste´matique et Evolution, CNRS UMR 8079, Universite´ Paris-Sud, Baˆtiment 360, 91405 Orsay Cedex, France

The dinophysoid dinoflagellates are currently Abbreviations: bp, base pairs; BV, bootstrap value; divided into three families: Amphisoleniaceae, Dino- s.s., sensu stricto physaceae (mainly Dinophysis Ehrenb. and Phalacro- ma F. Stein), and Oxyphysaceae, the latter including only one member, Oxyphysis oxytoxoides Kof. Phalac- The dinophysoids (see Appendix S1 in the sup- roma has been recently reinstated separately from plementary material for a note on the spelling of Dinophysis, and its amended description is currently the supergeneric names derived from Dinophysis restricted to cells whose epithecae were large but authored by P. C. Silva) are a well-defined order of <1 ⁄ 4 of the cell length. With the aim of improving marine dinoflagellates with 280 recognized the phylogeny of Dinophysales, we obtained 54 new classified in three families: Amphisoleniaceae, Dino- SSU rRNA gene sequences of 28 species. Taxon-rich physaceae, and Oxyphysaceae (Fensome et al. 1993, SSU rDNA phylogenetic analysis showed that Dino- Steidinger and Tangen 1997, Go´mez 2005). The physales split into two major clades, one containing cells are laterally compressed with a reduced epit- the Amphisoleniaceae (Amphisolenia F. Stein–Tripo- heca and a larger hypotheca consisting of two large solenia Kof.) and the other containing the Dinophys- plates united by a sagittal serrate suture with a zig- aceae. The latter are divided into two well- zag course (Kofoid and Skogsberg 1928, Tai and supported sister groups, the Dinophysaceae sensu Skogsberg 1934, Abe´ 1967a,b,c, Balech 1967). stricto (s.s.) (Dinophysis, F. Stein, Histi- According to Balech (1980), the dinophysoids are oneis F. Stein) and, tentatively, a separate family for also unusual among the thecate dinoflagellates in the clade of the type and most of the Phalacroma that, despite their extreme morphological specializa- species. Based on combined phylogenies of new tion, their plate arrangement and number are more SSU rDNA and available LSU rDNA data, O. oxytoxo- or less similar in all species and genera, except for ides (elongated epitheca, >1 ⁄ 4 of the cell length) the genera Amphisolenia F. Stein and Citharistes F. branched with a strong support with the type of Stein. Phalacroma. We therefore propose Phalacroma oxy- The dinophysoid genus Phalacroma F. Stein was toxoides comb. nov. for O. oxytoxoides. Our SSU morphologically separated from Dinophysis Ehrenb. rDNA phylogeny also suggests that the assumed high based mainly on differences in epithecal elevation intraspecific variability of Dinophysis hastata F. Stein (Stein 1883, Kofoid and Skogsberg 1928). Dinophysis hides a number of cryptic species. According to species have a reduced epitheca, and their anterior their distinct phylogenetic placement, the forms cingular list forms a funnel-shaped fan, whereas D. hastata f. phalacromides Jørg. and D. hastata Phalacroma species have a visible epitheca above an f. uracanthides Jørg. should be erected at the species anterior cingular list that is typically narrow and level. We propose for them the names Dinophysis directed horizontally. However, detailed tabulation phalacromoides comb. nov. and Dinophysis uracan- studies, especially at the sulcus level, did not show thoides comb. nov. any significant difference between Dinophysis and Key index words: Amphisolenia; cryptic species; Phalacroma. For this reason and because of their Dinophysiaceae; Dinophysiales; dinophysioid intergrading morphology, Abe´ (1967b) and Balech dinoflagellate; ; Ornithocercus; Oxyphysi- (1967) transferred Phalacroma species into Dinophy- aceae; Phalacroma; SSU and LSU rDNA phylogeny sis. More recently, and based on molecular data, Handy et al. (2009) and Hastrup Jensen and Daugb- jerg (2009) observed a deep separation between spe- 1Received 21 March 2010. Accepted 6 October 2010. cies of the Phalacroma and Dinophysis lineages. 2 Author for correspondence: e-mail fernando.gomez@fitoplancton. Accordingly, Hastrup Jensen and Daugbjerg (2009) com.

393 394 FERNANDO GO´ MEZ ET AL. reinstated the genus Phalacroma and amended its chambers at ·100 magnification with an inverted microscope description, which is currently restricted to cells (Nikon Eclipse TE200, Nikon Inc., Tokyo, Japan) and was whose epithecae are large but having <1 ⁄ 4 of the photographed at ·200 or ·400 magnification with a digital camera (Nikon Coolpix E995). Sampling continued from cell length. This definition excluded some Phalacro- October 2008 to August 2009 in the surface waters of the port ma species with a prominent epitheca (Phalacroma (depth of 2 m) of Banyuls sur Mer, France (4228¢50¢¢ N, apicatum Kof. et Skogsb. and Phalacroma cf. argus F. 308¢09¢¢ E), and from September 2009 to February 2010 in the Stein) that branched within the other Dinophysa- Bay of Villefranche sur Mer, Ligurian Sea. For the latter ceae clade (Dinophysis, Ornithocercus, Histioneis, location, sampling was performed at a long-term monitoring Citharistes). Hastrup Jensen and Daugbjerg (2009) site called Point B (4341¢10¢¢ N, 719¢00¢¢ E; water column divided the Dinophysales into three major clades: depth 80 m) and sporadically at Point C (4340¢ N, 719¢00¢¢ E; water column depth 600 m). Sampling in double one basal clade for Amphisolenia; another clade for oblique angle was performed with a custom-made conical species with the classical Phalacroma outline, includ- phytoplankton net (53 lm mesh size, 54 cm diameter and ing the type species of Phalacroma; and a third clade 280 cm length). The samples were prepared with the same with a variety of subclades (Ornithocercus and Cithar- procedure described above. The specimens were observed with istes, Histioneis, some Phalacroma species, and several an inverted microscope (Olympus IX51; Olympus Inc., Tokyo, Dinophysis clusters). These authors reported that Japan) and photographed with an Olympus DP71 digital camera. The specimen of O. oxytoxoides was collected by using Dinophysis was polyphyletic, as it formed four sepa- a strainer with netting of 20 lm aperture from the surface rate clades. This indicated that besides Dinophysis waters of the E´tang de Thau at Se`te (4324¢48¢¢ N, 341¢11¢¢ E) s.s., Dinophysis should be divided into at least three and analyzed following the procedure described above. E´tang additional new genera. de Thau is a large semienclosed coastal lagoon on the French These molecular phylogenetic studies still sup- Mediterranean coast (75 km2, mean depth 4.5 m) under the ported the division of the Dinophysales into three influence of freshwater inputs. In addition, open-water samples were collected from the BOUM (Biogeochemistry from the families: Amphisoleniaceae, Dinophysaceae, and Oligotrophic to the Ultra-oligotrophic Mediterranean) cruise Oxyphysaceae (Hastrup Jensen and Daugbjerg in the Mediterranean Sea between the Gulf of Lions and 2009). The latter is restricted to the monotypic Cyprus. Ten liters was collected from the surface with a bucket genus Oxyphysis Kof., which has a peridinioid and filtered by using a strainer of 20 lm netting aperture. The appearance, suggesting a possible link between din- retained material was fixed with absolute ethanol to a final ophysoids and peridinioids (Kofoid 1926). Nonethe- concentration of 50% concentrate seawater sample and 50% less, based on morphological traits, Abe´ (1967a) ethanol. At the laboratory, the ethanol sample was examined following the procedure described above. In all cases, each suggested that Oxyphysis might have a common specimen was micropipetted individually with a fine capillary ancestor with the Amphisolenia-Triposolenia line, leav- into a clean chamber and washed several times in serial drops ing its phylogenetic relationship unresolved. of 0.2 lm filtered and sterilized seawater (live specimens from With the aim of improving the phylogeny of coastal waters) or ethanol (ethanol prefixed specimens from Dinophysales, we obtained 54 new SSU rRNA gene open waters). Finally, the specimen was deposited into a sequences of 28 dinophysoid species from several 0.2 mL tubes (ABgene; Thermo Fisher Scientific Inc., Cour- taboeuf, France) filled with several drops of absolute ethanol. dinophysoid genera, including Amphisolenia, Tripo- The sample was kept at room temperature and in darkness solenia, and Oxyphysis as well as the type of Phalacro- until the molecular analysis could be performed. ma. Their phylogenetic analysis together with that of PCR amplification of SSU rRNA genes and sequencing. The combined available LSU rDNAs showed that specimens fixed in ethanol were centrifuged (Eppendorf, Dinophysales split into two major clades, the Amph- Hamburg, Germany) gently for 5 min at 504g. Ethanol was isoleniaceae (Amphisolenia-Triposolenia) and the then evaporated in a vacuum desiccator and single cells were resuspended directly in 25 lL of Ex TaKaRa buffer (TaKaRa, Dinophysaceae, the latter including two well-sup- distributed by Lonza Cia., Levallois-Perret, France). PCRs were ported groups, the Dinophysaceae s.s. (Dinophysis, performed in a volume of 30–50 lL reaction mix containing Ornithocercus, Histioneis) and Phalacroma species. 10–20 pmol of the eukaryotic-specific SSU rDNA primers EK- O. oxytoxoides branched within the Phalacroma, sup- 42F (5¢-CTCAARGAYTAAGCCATGCA-3¢) and EK-1520R (5¢- porting its transfer to this genus. D. hastata F. Stein CYGCAGGTTCACCTAC-3¢) (Lo´pez-Garcı´a et al. 2001). PCRs hides a number of cryptic species, and we propose were performed under the following conditions: 2 min dena- two of these be renamed D. phalacromoides comb. turation at 94C; 10 cycles of ‘‘touch-down’’ PCR (denaturation at 94C for 15 s; a 30 s annealing step at decreasing temper- nov. and D. uracanthoides comb. nov. ature from 65C down to 55C employing a 1C decrease with each cycle, extension at 72C for 2 min); 20 additional cycles at MATERIALS AND METHODS 55C annealing temperature; and a final elongation step of 7 min at 72C. A nested PCR was then carried out using 2–5 lL Sampling and isolation of material. The specimens were of the first PCR products in a GoTaq (Promega, Lyon, France) collected by slowly filtering surface seawater taken from the polymerase reaction mix containing the eukaryotic-specific end of the pier (depth 3 m) of the Station Marine d’Endoume, primers EK-82F (5¢-GAAACTGCGAATGGCTC-3¢) and EK- Marseille (4316¢48¢¢ N, 520¢57¢¢ E), from October 2007 to 1498R (5¢-CACCTACGGAAACCTTGTTA-3¢) (Lo´pez-Garcı´a September 2008. A strainer with netting of 20, 40, or 60 lm et al. 2001) and similar PCR conditions as described above. A mesh size (Millipore Inc., St. Quentin-Yveline, France) was used third, seminested, PCR was carried out using the dinoflagellate to collect the organisms, and the filtered volume varied specific primer DIN464F (5¢-TAACAATACAGGGCATCCAT-3¢) between 10 and 100 L, according to the concentration of (Go´mez et al. 2009) and keeping the reverse primer EK-1498R. particles. The concentrated sample was examined in Utermo¨hl MOLECULAR PHYLOGENY OF DINOPHYSALES 395

Negative controls without template DNA were used at all positions to be considered in our analysis. Similarly, from the amplification steps. Amplicons of the expected size 26 new SSU rDNA sequences of identified cells that these (1,200 base pairs [bp]) were then sequenced bidirectionally authors obtained, we only included the 11 sequences that were with primers DIN464F and EK-1498R using an automated 96- longer than 1,300 bp. To compare the position of O. oxytoxoides capillary sequencer ABI PRISM 3730xl (Cogenics, Meylan, in SSU and LSU rDNA phylogenies, we retrieved available LSU France). rDNA sequences of dinophysoid dinoflagellates from GenBank Phylogenetic analyses. The new SSU rDNA sequences were using Entrez (http://www.ncbi.nlm.nih.gov/sites/gquery) with aligned to a large multiple sequence alignment containing a taxonomic query. We reconstructed ML phylogenetic trees 1,100 publicly available complete or nearly complete for the LSU rDNA sequences alone or combined with SSU (>1,300 bp) dinoflagellate sequences using the profile align- rDNA sequences for the same species using TREEFINDER and ment option of MUSCLE 3.7 (Edgar 2004). The resulting the same substitution model specifications as for the SSU rDNA alignment was manually inspected using the program ED of the phylogenetic analysis (see above). Our sequences were depos- MUST package (Philippe 1993). Ambiguously aligned regions ited in GenBank under accession numbers HM853763– and gaps were excluded in phylogenetic analyses. Preliminary HM853816 (see Table S1 in the supplementary material). phylogenetic trees with all sequences were constructed using the neighbor-joining method (Saitou and Nei 1987) imple- mented in the MUST package (Philippe 1993). These trees RESULTS AND DISCUSSION allowed identifying the closest relatives of our sequences together with a sample of other dinoflagellate species, which Species identification. To amplify and sequence were selected to carry out more computationally intensive SSU rRNA genes from key dinophysoid species and maximum-likelihood (ML) analyses. These were done with the carry out phylogenetic analyses, we collected individ- program TREEFINDER (Jobb et al. 2004) applying a ual cells of a total of 28 dinophysoid species (Table GTR + C + I model of nucleotide substitution, taking into S1). All cells were individually identified, photo- account a proportion of invariable sites and a C-shaped graphed, and collected under the microscope distribution of substitution rates with four rate categories. Bootstrap values (BVs) were calculated using 1,000 pseudore- (Figs. 1–4). In the description of the specimens, we plicates with the same substitution model. follow the classification into three families accord- The phylogenetic position of the dinophysoids was analyzed ing to the current taxonomic scheme. by means of a global alignment of 77 taxa representing Family Amphisoleniaceae. This family is composed of sequences of dinoflagellates, including sequences of dinophy- two distinctive genera, Amphisolenia and Triposolenia, soid species, with representatives of the lineages of the characterized by an elongated cell body. We obtained , , and . We did not new SSU rDNA sequences for several species of include the environmental dinophysoid sequences of Handy et al. (2009) in our final SSU rDNA phylogenetic trees, as they the genus Amphisolenia, including the sequence of were short and imposed a limitation in the final number of the type, Amphisolenia globifera F. Stein, as well as

FIG. 1. Light micrographs of specimens of Amphisolenia and Triposolenia collected for single- cell PCR analysis. See Table S1 (in the supplementary material) for the collection date, location, and accession numbers. (a–b) Amphi- solenia globifera FG1401. (c) Amphi- solenia schauinslandii FG1163. (d) Amphisolenia sp. FG281. (e–f) Am- phisolenia bidentata FG279. (g–i) Other specimen of Amphisolenia bidentata under epifluorescence microscopy. Note the intracellular coccoid symbionts. (j) Triposolenia bicornis 2 FG1153. (k) T. bicornis 4 FG1155. (l) T. bicornis 6 FG1157. (m) T. bicornis 8 FG1159. Scale bars, 20 lm. 396 FERNANDO GO´ MEZ ET AL.

FIG. 2. Light micrographs of Phalacroma. See Table S1 (in the supplementary material) for the collection date, location, and accession numbers of the specimens collected for single-cell PCR analysis. (a) Phalacroma porodictyum FG487. (b–c) P. porodictyum FG510. (d–e) P. porodictyum FG519. (f–g, i) P. porodictyum FG490. (h) Original illustration of P. porodictyum by Stein (1883). (j–m) Live specimen of P. porodictyum FG1193b. (j) Note the pores in the . (l) The arrows indicate the two rows of pores along the cingulum. (n) Phalacroma sp. FG517. (o) Phalacroma mitra FG175. (p) P. mitra FG175b. (q) P. mitra FG525. (r) P. mitra FG1179. (s–t) Phalacroma rapa FG1187. (u) Phalacroma favus FG1183. (v–w) P. favus FG1188. The arrows (t, u) indicate the different length of the third rib. (x) Phalacroma doryphorum FG365. (y) P. doryphorum FG641. (z) P. doryphorum FG509. (aa) Phalacroma rotundatum FG366. (ab) P. rotundatum FG365. (ac–ad) Phalacroma parvulum FG503. (ae–af) P. parvulum FG505. (ag) P. parvulum FG326. Scale bars, 20 lm. for the genus Triposolenia. A. globifera, one of the (765 lm long, 21 lm wide) contained green granules smaller species of the genus, is characterized by a in the central body and was identified as Amphisolenia swelling of the antapical end. The specimen studied bidentata Schro¨d. (Fig. 1, e and f). In other similar (207 lm long, 12 lm wide) showed a swelling of specimens, these green granules appeared to contain 8 lm in diameter with two small spines (Fig. 1, a and chl a, as seen under epifluorescence microscopy b). Amphisolenia schauinslandii Lemmerm., closely (Fig. 1, g–i). related to the type, was larger and showed an inflated While Amphisolenia, especially A. bidentata, was midbody (390 lm long, 35 lm wide) (Fig. 1c). common in surface waters, the specimens of Tripo- Another specimen (335 lm long, 19 lm wide; solenia were preferentially distributed in deep waters. Fig. 1d) resembled Amphisolenia complanata Kof. et We obtained identical SSU rDNA sequences from Skogsb., but it might represent an incompletely devel- four live specimens of Triposolenia bicornis Kof. from oped specimen and was therefore more difficult to the same sample collected at 400 m depth. The identify. To avoid misnaming, we generically named it dimensions of the four cells were identical, with a Amphisolenia sp. The largest Amphisolenia specimen length from the apex to the tip of the antapical MOLECULAR PHYLOGENY OF DINOPHYSALES 397

FIG. 3. Light micrographs of Dinophysis specimens collected for single-cell PCR analysis and other live or Lugol’s-fixed specimens. See Table S1 (in the supplementary material) for the collection date, location, and accession numbers of the sequenced specimens. (a) Din- ophysis caudata FG178. (b) Dinophysis tripos FG56. (c–d) Dinophysis hastata FG1432. (e) Other specimen of D. hastata from the same sample. (f) Note the areolation of the empty theca of D. hastata. (g–h) D. hastata f. uracanthides FG499. (i) D. hastata f. uracanthides FG527. (j) Ori- ginal illustration of D. hastata reproduced from Stein (1883). (k–l) First and second illustration of Dinophysis uracantha in Stein (1883). (m) Dinophysis swezyae in Kofoid and Skogsberg (1928). (n) D. hastata f. uracanthides in Jørgensen (1923). (o) Dinophysis balechii in Norris and Berner (1970). (p) Dinophysis alata in Jørgensen (1923). (q) Dinophysis uracantha var. mediterranea in Jørgensen (1923). (r) D. hastata f. phalacromides in Jørgensen (1923). (s) Dinophysis odiosa in Pavillard (1930). (t) Dinophysis monacantha in Kofoid and Skogsberg (1928). (u) D. uracantha in Jørgensen (1923). (v) Dinophysis pusilla in Jørgensen (1923). (w) Dinophysis schuettii (smaller form) in Jørgensen (1923). (x) Dinophysis acutissima in Gaarder (1954). (y) Dinophysis reticulata in Gaarder (1954). (z–aa) D. hastata f. phalacromides FG1170. (ab) Din- ophysis odiosa FG176. (ac–ad) D. odiosa FG1429. (ae) Another specimen of D. odiosa. (af–ag) Dinophysis monacantha FG1414. (ah) Ethanol- fixed specimen of D. pusilla FG497. (ai) Ethanol-fixed specimen of D. pusilla FG497. (aj) Live specimen of D. cf. pusilla FG524. (ak) Lu- gol’s-fixed specimen from the NW Pacific Ocean with morphology between D. pusilla and D. balechii. (al) D. uracantha var. mediterranea. The inset shows the two lateral ribs in the antapical spine. (am) D. uracantha. (an) D. schuettii. (ao) Ethanol-fixed specimen of Dinophysis cf. acutissima FG523. (ap) Live specimen of D. acutissima. Scale bars, 20 lm. extensions of 150 lm, the basis of cell body of sequences for the type, Phalacroma porodictyum F. 45 lm, and the diameter of the neck or the antapi- Stein, and Phalacroma favus Kof. et J.R. Michener, cal horns of 4 lm. The cell body shapes showed Phalacroma parvulum (F. Schu¨tt) Jørg., Phalacroma slight differences among the specimens, from trian- mitra F. Schu¨tt, and Phalacroma doryphorum F. Stein, gular to more rotund contours (Fig. 1, j–m). with 2–5 sequences for each species. We also deter- Family Dinophysaceae. For the description of the mined the sequence of a nonidentified Phalacroma specimens of the genera Phalacroma and Dinophysis, species. we follow the classification into sections proposed Section Euphalacroma Jørg.: This section included by Pavillard (1916) and Jørgensen (1923). the genus type as illustrated by Stein (1883) (Fig. 2h). Genus Phalacroma: The SSU rDNA sequences The contour of the P. porodictyum cells collected in identified at the species level available in GenBank our samples was slightly oval (66 lm long, 62 lm are limited to Phalacroma rotundatum (Clap. et wide) with a dome-shaped epitheca clearly projecting Lachm.) Kof. et J.R. Michener and Phalacroma rapa over the margin of the upper cingular list (Fig. 2, F. Stein. In addition to several additional sequences a–m). The cells were apochlorotic, and the theca of these species, we obtained new SSU rDNA showed regularly scattered pores and two rows of 398 FERNANDO GO´ MEZ ET AL.

FIG. 4. Light micrographs of Ornithocercus, Histioneis, and Oxy- physis collected for single-cell PCR analysis. See Table S1 (in the supplementary material) for the collection date, location, and ac- cession numbers. (a) Ornithocercus magnificus FG25 (b) Ornithocercus heteroporus FG324. (c) O. heteroporus FG323. (d) O. heteroporus FG507. (e) O. heteroporus FG506. (f) Orni- thocercus quadratus var. quadratus FG1004. (g) O. quadratus var. quad- ratus FG1174. (h) O. quadratus var. schuettii FG1173. (i) H. cymbalaria FG325. (j) Histioneis longicollis FG1167. (k) H. longicollis FG1168. (l) Histioneis gubernans FG26. (m) Another specimen of H. gubernans. (n) Oxyphysis oxytoxoides FG278. (o) Lugol’s-fixed specimen of an undescribed dinophysoid from the Pacific Ocean. Note the elon- gate epitheca. Scale bars, 20 lm.

pores along the cingulum (Fig. 2l). The list between (Fig. 2, o–r). There is a historical controversy on the first and second ribs was covered with an irregular the synonymy of P. rapa and P. mitra (Schiller reticulum, while the region between the second and 1933), but we observed that P. rapa cells were larger third ribs was smooth. In the ventral side of the hy- (61 lm long, 83 lm wide) and exhibited a greater potheca, there was a linear structure that connected angularity of the ventral margin than those the second rib and the cingulum (Fig. 2, a and d). of P. mitra when seen in lateral view (Fig. 2, s and We obtained five SSU rDNA sequences from live and t). P. favus (64 lm long, 86 lm wide) differed from ethanol-fixed specimens from the coastal and open P. rapa principally in the constricted, projecting Mediterranean waters (Table S1). fingerlike antapex. The length of the third rib of Unclassified Phalacroma: We were unable to iden- P. favus was smaller than in P. rapa (Fig. 2, u–w). tify one Phalacroma specimen at the species level Section Urophalacroma Jørg.: We collected cells because information on the left sulcal list was diffi- belonging to the type of this section, P. doryphorum, cult to obtain. The cell was slightly elliptical (46 lm which had a wedge-shaped hypotheca, dome-shaped long, 40 lm wide) with a prominent epitheca of epitheca, and distinct horizontal cingular list. The the same width as the hypotheca. The general left sulcal list was well-developed, and the most dis- appearance resembled that of Phalacroma ovum tinctive character was a nonribbed wide posterior Schu¨tt, but to avoid a possible misnaming, we called projection with a triangular shape (Fig. 2, x–z). it Phalacroma sp. (Fig. 2n). Section Paradinophysis Jørg.: We obtained SSU Section Podophalacroma Jørg.: Described species rDNA sequences of two specimens of P. rotundatum of this section are characterized by an asymmetrical collected from the same sample. The cell contour wedge-shaped hypotheca with a prominent, large was ellipsoidal, the theca was smooth, and the epit- areolation in the theca and a greenish pigmenta- heca was hardly visible above the cingulum. One of tion. The diagnostic criteria for the species differen- the specimens was slightly larger and had a wider tiation are the shape and size of the hypotheca. We hypotheca than the other one (Fig. 2, aa and ab). observed different species from this section in our We also obtained sequences of three specimens of samples. P. mitra specimens (58 lm long, 46 lm P. parvulum from open and coastal waters (Table wide) had a distinctive broad wedge-shaped hypot- S1). The cells (40 lm long, 33 lm wide) had a heca. The dorsal side was convex, and the ventral smooth theca and showed a regularly round outline side was more or less straight in the sulcus region, in lateral view. The epitheca was dome-shaped with becoming distinctly concave at the posterior end of horizontal cingular lists (Fig. 2, ac–ag). the left sulcal list toward the antapical end. The Genus Dinophysis: Various sequences of the chlo- epitheca was almost flat with horizontal cingular lists roplast-containing species of Dinophysis s.s. were MOLECULAR PHYLOGENY OF DINOPHYSALES 399 available in GenBank. However, no complete SSU forms of D. hastata and D. uracantha from the open rDNA sequence of any apochlorotic species of Mediterranean Sea. Kofoid and Skogsberg (1928) Dinophysis was available. In addition to several new described two species with antapical spines, D. mona- sequences of members of Dinophysis s. s., we deter- cantha (Fig. 3t) and D. urceolus Kof. et Skogsb., which mined SSU rDNA sequences of several apochlo- were further considered as synonyms of D. hastata rotic species, including different morphotypes of (Abe´ 1967b). Kofoid and Skogsberg (1928) proposed D. hastata, as well as Dinophysis odiosa (Pavill.) L. S. the existence of a high intraspecific variation in Tai et Skogsb., Dinophysis monacantha Kof. et D. hastata. Pavillard (1930) remarked the validity of Skogsb., Dinophysis pusilla Jørg., and Dinophysis cf. D. odiosa (Fig. 3s) that was synonymized with D. hasta- acutissima Gaarder. ta (Kofoid and Skogsberg 1928, Taylor 1976). Norris Section Homoculus Pavill.: We determined the and Berner (1970) described Dinophysis balechii D. R. complete SSU rDNA of Dinophysis tripos Gourret. Norris et L. D. Berner (Fig. 3o) previously considered The specimen used was collected from its type local- one of the small forms of D. hastata. ity, the Bay of Marseille (Fig. 3a). In addition, we To clarify the phylogenetic position of D. hastata obtained an additional sequence for Dinophysis and other members of this section, we sampled caudata Kent (Fig. 3b). specimens of the section Hastata that we classified Section Hastata Pavill.: The species of this section into two different subsections according to their are apochlorotic and characterized by antapical morphology: specimens with a Dinophysis-like epit- spines. We illustrated the single-cell sequenced spec- heca and funnel were considered as ‘‘uracantho- imens and other specimens to show the differences ides,’’ and specimens with flat epitheca and among the species, and we reproduced the original cingular list with a Phalacroma-like cingular list were illustrations of the species and varieties related to included in the subgroup ‘‘phalacromoides.’’ D. hastata found in Stein (1883), Jørgensen (1923), Subsection uracanthoides: This group contained Kofoid and Skogsberg (1928), Pavillard (1930), the type of the section, D. hastata. The specimens of Gaarder (1954), and Norris and Berner (1970) to this group have a morphology intermediate between provide a reference for the accuracy of our speci- the original description of D. hastata (Fig. 3j) and men identifications (Fig. 3, j–y). Due to the histori- the second of Stein’s illustration of D. uracantha cal controversy about the synonymy and supposed (Fig. 3l). We suspect that Stein might have exces- intraspecific variability of some species of the sec- sively elongated the cell body of D. hastata in his draw- tion Hastata, especially the type D. hastata,we ing. The members of uracanthoides were characterized extend the description of this section below. by an elliptical cell body and a small epitheca with a The description of the first member of the sec- funnel-shaped cingular list as in typical Dinophysis s.s. tion Hastata, D. hastata, is credited to Stein (1883, The third rib of the left sulcal list emerged from the pl. 19, fig. 12), who provided a single illustration of lower half of the hypotheca. We obtained SSU rDNA D. hastata (Fig. 3j). Since then, and although the sequences for two morphotypes of this kind. The records under the name D. hastata were numerous, first morphotype was observed in live specimens that specimens with the same hypotheca contour as in were 64 lm long and 51 lm wide (Fig. 3, c–f), with a Stein’s original description were never observed. All dorsoventral diameter at the base of the funnel authors citing D. hastata assumed that the hypotheca (upper girdle list) of 27 lm. Our specimens strongly of D. hastata was rounder than the Stein’s specimen. resembled the original illustration of D. hastata Stein (1883) also described Dinophysis uracantha F. for the antapical spine and coarse left sulcal reticu- Stein, another Dinophysis species with an antapical lation (Fig. 3f), and, therefore, we ascribed them to spine. However, under the name D. uracantha, Stein D. hastata (Fig. 3, c–f). The second morphotype (1883, pl. 20, figs. 21 and 22) provided two illustra- resembled the description of D. hastata f. uracanthides tions that unequivocally corresponded to two differ- Jørg. described by Jørgensen (1923) from the Medi- ent species based on current morphological criteria terranean Sea (Fig. 3n), for which D. hastata (Fig. 3, k and l). The first illustration of D. uracantha f. uracanthides was smaller than D. hastata and had a (Fig. 3k) was very similar to specimens of D. uracantha narrower and more ventrally deflected antapical observed in this study (59 lm length, 49 lm wide) spine. In contrast to the specimens of D. hastata, the (Fig. 3am), close to Dinophysis swezyae Kof. et Skogsb. third rib in the D. hastata f. uracanthides specimens (Fig. 3m). The second illustration of D. uracantha that we observed did not reach the level of the basis (Fig. 3l) showed a specimen with a larger and ovate of the hypotheca (Fig. 3, g–i). We collected two speci- hypotheca, a very large antapical spine, and a long mens with these characteristics from the same station third rib extending below the basis of the epitheca. in the Levantine Basin, which were ethanol-fixed This Stein’s second illustration of D. uracantha has for posterior SSU rDNA amplification and sequen- been considered as synonym of D. hastata (Abe´ cing (Table S1). Their cell body was ellipsoidal 1967b). Both Stein’s illustrations of D. uracantha also (46 lm long, 34 lm wide), and the dorsoventral differed in the reticulation, which was coarser for the diameter of the base of the ‘‘funnel’’ (upper girdle illustration used to consider D. uracantha as synonym list) was 21 lm (Fig. 3, g–i). Other members of this of D. hastata. Jørgensen (1923) described several subsection are D. balechii (Fig. 3o), D. uracantha var. 400 FERNANDO GO´ MEZ ET AL. mediterranea Jørg. (Fig. 3, q and al), Dinophysis alata well-developed left sulcal list (Fig. 3v). The antapical Jørg. (Fig. 3p), and Dinophysis spinosa Rampi. spine with a prominent rib was slightly ventrally Subsection phalacromoides: The members of deflected and emerged from the posterior-ventral phalacromoides are characterized by a flat and wider region of the hypotheca. We obtained SSU rDNA epitheca. The upper cingular list slightly extended sequences from two ethanol-fixed specimens col- over the epitheca. The funnel-shaped upper cingu- lected in the open Mediterranean Sea. Both were lar list of previous specimens was lacking, resem- 30 lm wide, with a dorsal-ventral diameter of the bling that of Phalacroma. Species of this subgroup funnel base of 14 lm. One of the specimens showed include D. odiosa, first described as Phalacroma a rotund hypotheca, with a straight third rib odiosum Pavill. As a general trend, the specimens of (Fig. 3ah). This morphology unequivocally corre- phalacromoides were larger and the hypotheca was sponded to D. pusilla. The second specimen was more ovate that in members of uracanthoides. The slightly larger, with an ovate contour of the epitheca third rib of the sulcal list emerged from the middle and the third rib of the sulcal list deflected posteri- of the ventral side of the hypotheca and did not orly. A single prominent rib was projected anteriorly extend beyond the basis of the epitheca. The ant- from the cingulum (Fig. 3ai). We also provided the apical spine was always ventrally deflected. We iden- illustration of a live specimen (28 lm long, 26 lm tified three morphotypes having these wide, diameter of funnel base 8 lm) (Fig. 3aj) and morphological characteristics, those corresponding of a Lugol’s-fixed specimen from the Pacific Ocean to D. odiosa, D. hastata f. phalacromides Jørg., and with a morphology that was intermediate between D. monacantha. D. pusilla (Fig. 3v) and D. balechii (Fig. 3o). We D. odiosa is one of the most common species of ascribed the species Dinophysis schuettii G. Murray et the section Hastata in the coastal Mediterranean Whitting (Fig. 3, w and an) to this group. Sea. However, it is rarely cited in the Mediterranean Subsection acutissima: Gaarder (1954) described Sea (Go´mez 2003), very likely because it is incor- two species, D. acutissima (Fig. 3x) and Dinophysis rectly reported as D. hastata. Its cell body was some- reticulata Gaarder (Fig. 3y), characterized by an what truncate anteriorly, fairly narrowly rounded elongated hypotheca with a pronounced antapex posteriorly (Fig. 3, ab–ae), while the cell contour of that resembled the morphology of Dinophysis D. hastata was ovoid (Fig. 3, c–f). The epitheca of D. diegensis Kof. The distinctive character of this species odiosa was flat, wider than the greatest height of was a short antapical spine. Species with these char- epitheca. The anterior cingular list showed numer- acteristics were not included in the sections estab- ous radial ribs. The third rib was longer and more lished by Pavillard (1916) and Jørgensen (1923). posteriorly deflected in D. hastata than in D. odiosa Norris and Berner (1970) reported D. reticulata (Fig. 3, ab–ae). We obtained sequences from two among the members of the D. hastata group, but specimens of D. odiosa from different locations they did not mention D. acutissima. To the best of (Marseille and Villefranche sur Mer). They were our knowledge, the only record after the first 75 lm long and 59 lm wide, with a dorsoventral description corresponded to Nguyen et al. (2008). diameter of the base of the upper cingular list of These authors placed D. acutissima in the group 50 lm (Fig. 3, ab–ad). The second morphotype cor- doryphorum with P. doryphorum as type. We disagree responded to Dinophysis hastata f. phalacromides Jørg. with this view, and we preferred to place D. acutiss- described from the Mediterranean Sea by Jørgensen ima as a member of the section Hastata. We illus- (1923) (Fig. 3r). We collected a live specimen that trated a live specimen of D. acutissima (60 lm long, was the largest observed for this section (79 lm 38 wide, diameter of funnel base 28 lm) (Fig. 3ap). long, 62 lm wide, with a dorsoventral diameter of As far as we know, this was the first record for the the base of the upper cingular list of 41 lm; Fig. 3, Mediterranean Sea (Go´mez 2003). We obtained the z and aa). The third morphotype of this subsection SSU rDNA sequence from one ethanol-fixed speci- corresponded to D. monacantha (Fig. 3t). The lack men from the open Ionian Sea (Table S1). This eth- of citations of this species is likely due to the fact anol-fixed specimen of D. acutissima was 60 lm long that it has been considered a synonym of D. hastata and 40 lm wide, and the dorsal-ventral diameter of (Abe´ 1967b) and consequently pooled as D. hastata. the funnel base was of 24 lm (Fig. 3ao). D. monacantha was the smallest species observed for Genus Ornithocercus: This ornamented hetero- this subsection (67 lm long, 50 lm wide, base of trophic genus is characterized by a highly developed the upper cingular list, 38 lm in diameter) and its cingular chamber formed by the cingular list, which general morphology resembled that of a small harbors unicellular . The left sulcal list D. odiosa with a less flat and wider epitheca (Fig. 3, of these species is also highly developed, with ribs af and ag). or keels that emerged from the hypotheca. We Subsection Pusilla: The members of this group obtained the SSU rDNA sequence of Ornithocercus contained the smallest species of the section magnificus F. Stein, the type species collected from Hastata. Jørgensen (1923) described D. pusilla as a the type locality, the NW Mediterranean Sea small species with rotund hypotheca (28 lm wide), (Fig. 4a). We also obtained sequences of four speci- prominent funnel-shaped upper cingular list, and a mens of Ornithocercus heteroporus Kof. from live MOLECULAR PHYLOGENY OF DINOPHYSALES 401

(Fig. 4, b and c; Table S1) and ethanol-fixed speci- branched within a large dinoflagellate group com- mens from different locations of the Mediterranean posed of taxa of the orders Gymnodiniales, Peridini- Sea (Fig. 4, d and e; Table S1) and of specimens of ales, and Prorocentrales. However, this relationship two varieties of Ornithocercus quadratus var. quadratus was poorly supported (BV of 62%). Within this Kof. et Skogsb. (Fig. 4, f and g) and var. schuettii group, symmetric species of Prorocentrum Ehrenb. Kof. et Skogsb. (Fig. 4h). (Prorocentrum lima Ehrenb., Prorocentrum concavum Genus Histioneis: Only one complete SSU rDNA Fukuyo, and P. levis M. A. Faust, Kibler, Vandersea, sequence of Histioneis was available in GenBank, P. A. Tester et Litaker) branched as sister group of derived from a specimen that was not identified to the dinophysoid clade, although without support the species level, illustrated in dorsoventral view in (BV <50%). Concerning the internal phylogeny of the original publication (Handy et al. 2009), which the Dinophysales, this general phylogenetic tree made difficult its posterior identification. We deter- provided good support for the sister relationship of mined four SSU rDNA sequences of three species the genera Amphisolenia and Triposolenia (BV of identified at the species level. The type of Histioneis, 88%), whereas the rest of species emerged within a Histioneis remora F. Stein, has been very scarcely strongly supported group (BV of 99%) (Fig. 5). recorded, and the few existing records are doubtful. To obtain a more resolved view of the phyloge- Stein’s illustration did not allow defining the type netic relationships among the Dinophysales, we con- species. Stein (1883) also provided under the name structed a SSU rDNA data set restricted to these Histioneis cymbalaria F. Stein three illustrations that species. The corresponding phylogenetic tree unequivocally corresponded to three separate spe- (Fig. 6) was rooted with the clade of Amphisolenia- cies. We obtained the sequence of a specimen ceae (according to the results of the previous gen- strongly resembling one of the Stein’s illustrations eral analysis containing other dinoflagellates, Fig. 5). of H. cymbalaria. Hence, although this species has The other major clade, containing all the remaining been named either Histioneis depressa J. Schiller by dinophysoids, was divided into two well-supported Taylor (1976) or H. cymbalaria by Balech (1988), large subclades: one for species with the classical our specimen has been ascribed to H. cymbalaria fol- Phalacroma morphology, which also included O. oxy- lowing Go´mez (2007) (Fig. 4i). We sequenced SSU toxoides (BV of 91%), and a second one for the spe- rDNAs from two specimens of Histioneis longicollis cies of Dinophysis, Ornithocercus, and Histioneis (BV of Kof. collected from the Bay of Villefranche sur Mer 100%). Our phylogenetic tree did not support the on two consecutive days (Fig. 4, j and k). Both speci- consideration of O. oxytoxoides as the only member of mens were identical in size (83 lm maximum its own family and even as a separate genus from length, and the width of the hypotheca was 28 lm) Phalacroma. On the contrary, it emerged well nested with a peculiar yellow-greenish brightness of the sul- among the Phalacroma species (BV of 81%), closely cal list, although they slightly differed in internal related to the type of Phalacroma (Fig. 6). ornamentation of the left sulcal list. The morphol- To test the robustness of this unexpected place- ogy corresponded to Histioneis sublongicollis Halim ment of O. oxytoxoides within the clade containing described from the Bay of Villefranche. Following the type of Phalacroma, we constructed an LSU Go´mez (2007) H. sublongicollis was considered a rDNA tree using the relatively large sampling of din- synonym of H. longicollis. The other sequenced ophysoid sequences available at GenBank (including specimen belonged to the Histioneis gubernans O. oxytoxoides: EF613359). The LSU rDNA phylogeny group (Go´mez 2007), closely related to Histioneis (Fig. S1 in the supplementary material) also sup- striata Kof. et J. R. Michener, and was ascribed to ported the late emergence of O. oxytoxoides among H. gubernans F. Schu¨tt (Fig. 4l). Another live speci- the Phalacroma species, with even stronger support men of H. gubernans is illustrated for comparison than the SSU rDNA (BV of 89%). Finally, we carried (Fig. 4m). out a combined phylogenetic analysis of concate- Family Oxyphysaceae. O. oxytoxoides is the only nated SSU and LSU rDNA sequences for the dino- member of this family. No specimen was observed physoid species for which sequences of both in the coastal or open Mediterranean Sea. We markers are available. The SSU + LSU rDNA tree obtained the SSU rDNA sequence from a specimen (Fig. 7) provided very strong support for the inclu- (58 lm long, 20 lm wide) collected in the brackish sion of O. oxytoxoides within the genus Phalacroma waters of the Thau lagoon at Se`te, south of France (BV of 100%), more precisely within a subclade (Fig. 4n). In addition to Oxyphysis, a nondescribed clustering the Phalacroma type, P. porodictyum, and P. species from the Pacific Ocean with an elongated doryphorum (BV of 100%). epitheca is shown (Fig. 4o). SSU and LSU rDNAs have very contrasted Molecular phylogeny. We constructed ML trees degrees of variability in the dinophysoid species. For from a global alignment of dinoflagellate SSU rDNA example, the sequence identity between O. oxytoxo- sequences using other as outgroup. All ides and P. porodictyum was 99.27% for the SSU the sequences of representative dinophysoid dinofla- rDNA and 93.82% for the LSU rDNA. Therefore, gellates emerged within a strongly supported (BV of the phylogenetic position of O. oxytoxoides as a mem- 94%) monophyletic clade (Fig. 5). This clade ber of the genus Phalacroma retrieved support from 402 FERNANDO GO´ MEZ ET AL.

FIG. 5. Maximum-likelihood phylogenetic tree of dinoflagel- late SSU rDNA sequences, based on 1,166 aligned positions. Names in bold represent sequences obtained in this study. Numbers at nodes are bootstrap values (values <50% are omitted). Accession numbers are provided between brackets. The scale bar represents the number of substitutions for a unit branch length.

markers with different levels of conservation, stress- P. parvulum, P. mitra, and P. rapa (BV of 55%). We ing its reliability. called the former Phalacroma subclade because it Internal SSU rDNA phylogeny of the three major contained the type, and the second Rapa because dinophysoid groups. Amphisoleniaceae: The species of P. rapa was the first described species in that subc- Amphisolenia and Triposolenia formed two highly sup- lade (Fig. 6). ported groups (BV of 86% and 100%, respectively). The classical subdivision into sections represented However, the relative branching order of the differ- by two or more species as Paradinophysis (P. rotunda- ent Amphisolenia species was weakly supported, tum, P. parvulum) and Podophalacroma (P. mitra, although the tree suggested that the two morpho- P. rapa, P. favus) is not supported by the molecular logically closely related species A. globifera and data, since species of same section appeared in dif- A. schauinslandii were sister (BV of 68%), with ferent subclades of Phalacroma. In fact, P. rotundatum A. bidentata branching in a basal position (Fig. 6). branched in the subclade Phalacroma, close to Phalacroma: Although several nodes within the P. porodictyum and O. oxytoxoides with relatively good Phalacroma clade were not well resolved, it support (BV of 86%), whereas P. parvulum branched appeared that this clade was subdivided into two in the subclade Rapa. P. mitra and P. rapa were sis- subclades: one for the type, P. porodictyum, and ters in the subclade Rapa (BV of 97%). Surprisingly, P. favus, P. rotundatum, P. doryphorum, and Oxyphysis despite the strong morphological resemblance, (BV of 81%); and a second one containing P. favus is distantly related to other members of MOLECULAR PHYLOGENY OF DINOPHYSALES 403

FIG. 6. Maximum-likelihood phylogenetic tree of Dinophysales SSU rDNA sequences, based on 1,166 aligned positions. Names in bold represent sequences obtained in this study. Numbers at nodes are bootstrap values (values <50% are omitted). The sequences of the type species are highlighted in gray shaded boxes. Accession numbers are provided between brackets. The scale bar represents the number of substitutions for a unit branch length.

Podophalacroma, and it branched closely related to tical. This extreme conservation of the SSU rDNA P. porodictyum in the Phalacroma subclade. within the genus Ornithocercus made this marker Dinophysaceae: Our phylogenetic analysis did not inappropriate to resolve the relationships among provide a robust resolution of all genera in the the corresponding species. Dinophysaceae clade, which showed large variations The sequences of Histioneis formed a group that of evolutionary rate, as depicted by the extreme appeared to be more closely related to Ornithocercus differences of branch lengths (Fig. 6). Ornithocercus than to Dinophysis. Histioneis cymbalaria emerged at and Histioneis had short branches and emerged in a the base of H. longicollis and H. gubernans (BV of basal position with respect to Dinophysis s.s., 67%). However, the poor internal resolution for this although with moderate support (BV of 68%). All genus made it premature to draw conclusions about the Ornithocercus sequences were almost completely the Histioneis systematics only on the basis of the identical, with the exception of O. magnificus FTL83 SSU rDNA phylogeny. (Handy et al. 2009), probably because of a few The -containing species of Dinophysis, sequence errors. Sequences of the two O. quadratus the so-called Dinophysis s.s. that contained the type varieties (var. schuettii and var. quadratus) were iden- species, formed a well-supported monophyletic 404 FERNANDO GO´ MEZ ET AL.

Nomenclatural considerations. The placement of O. oxytoxoides closely related to the type species of Phalac- roma in the SSU and LSU rDNA phylogenies supports the transfer of O. oxytoxoides into Phalacroma. Phalacroma oxytoxoides (Kof.) F. Go´mez, P. Lo´pez- Garcı´a et D. Moreira, comb. nov. Basionym: Oxyphysis oxytoxoides Kof. (Kofoid 1926, p. 205, pl. 18). The type of Dinophysis, Dinophysis acuta Ehrenb., formed a well-defined clade within the Dinophysis s.s. with other chloroplast-containing species of Din- ophysis. In contrast, the sequences of D. hastata f. phalacromides and D. hastata f. uracanthides, as well as those of other apochlorotic members of the Din- FIG. 7. Maximum-likelihood phylogenetic tree of Dinophysales LSU and SSU rDNA sequences, based on 1,980 aligned positions. ophysis hastata group, branched very distantly from Numbers at nodes are bootstrap values (values <50% are omit- the type of Dinophysis, which might support their ted). The sequences of the type species are highlighted in gray inclusion in different genera. However, the even- shaded boxes. Accession numbers are provided between brackets. tual split of Dinophysis and the erection of a new The scale bar represents the number of substitutions for a unit branch length. genus for the members of the D. hastata group are premature. Additional information from other molecular markers would be required to obtain more robust phylogenies that would eventually vali- clade. The other species of Dinophysis belonging to date this claim. Nevertheless, the molecular data the section Hastata, apochlorotic with an antapical clearly demonstrated that the assumed high intra- spine, branched in a basal position to the members specific variability of D. hastata hid a number of of Dinophysis s.s. Our molecular phylogeny strongly cryptic species. In fact, species such as D. odiosa or supports D. odiosa and D. monacantha as separate D. monacantha, previously considered as synonyms species, given the large divergence between their of D. hastata (Abe´ 1967b, Taylor 1976), appeared sequences (Fig. 6). Likewise, the sequences of the to be distant species on the basis of their high SSU two forms of D. hastata,f.phalacromides and f. ura- rDNA sequence divergence (Fig. 6). Likewise, our canthides, were very divergent from that of D. hasta- molecular phylogenetic analysis showed that the ta, so that these forms also deserve to be two forms D. hastata f. phalacromides and D. hastata considered as two separate species. The species of f. uracanthides described by Jørgensen (1923) do the section Hastata were relatively distant from correspond to separate phylogenetic species. Conse- Dinophysis s.s. in all the SSU rDNA phylogenies. quently, we propose to erect these forms at the spe- However, the internal relationships between the cies level as follows: members of this section were unstable and poorly Dinophysis phalacromoides (Jørg.) F. Go´mez, P. supported in different molecular phylogenies. This Lo´pez-Garcı´a et D. Moreira, comb. nov. trend is probably due, at least partly, to the Basionym: Dinophysis hastata F. Stein f. phalacro- extreme differences of evolutionary rate in these mides Jørg. (Jørgensen 1923, pp. 30–1, fig. 41). species, with short-branching ones such as D. pusilla Dinophysis uracanthoides (Jørg.) F. Go´mez, P. and D. cf. acutissima, and very long-branching ones Lo´pez-Garcı´a et D. Moreira, comb. nov. such as D. monacantha and D. odiosa. D. hastata f. Basionym: Dinophysis hastata F. Stein f. uracanthides uracanthides branched as sister of the Dinophysis s.s., Jørg. (Jørgensen 1923, pp. 30–1, fig. 40). making the Hastata paraphyletic (Fig. 6). Neverthe- Reconciling morphological and molecular data. The less, this finding was weakly supported (BV of detailed studies on the tabulation carried out by Tai 62%), and we cannot exclude that the paraphyly of and Skogsberg (1934), Abe´ (1967a,b,c), Balech the Hastata reflected a long-branch attraction arti- (1967, 1988), and Norris and Berner (1970) showed fact due to the long branches of certain members. that the plate arrangement and number are more The only well-supported subgroup is composed of or less similar in all dinophysoid species and genera, D. odiosa, D. monacantha, D. pusilla, Dinophysis cf. even at the level of the small sulcal plates. Abe´ acutissima, and D. hastata f. phalacromides (BV of (1967a) observed some differences in the tabulation 97%) (Fig. 6). Within this group, there was a very for the genera Amphisolenia, Triposolenia, and Oxyphy- short distance between Dinophysis cf. acutissima, with sis when compared with the other dinophysoids. All an elongated and antapically pointed hypotheca, authors agreed on the establishment of the family and D. pusilla. This relationship suggested that the Amphisoleniaceae for Amphisolenia and Triposolenia shape of the hypotheca is a morphological charac- (Abe´ 1967a, Balech 1977, 1980). Our molecular phy- ter highly variable among very closely related spe- logenetic analysis including SSU rDNA sequences of cies and, consequently, of relatively small value as a Triposolenia and the type of Amphisolenia, which phylogenetic marker of the species. appear as monophyletic and occupy a basal position MOLECULAR PHYLOGENY OF DINOPHYSALES 405 within the Dinophysales, supports their consideration species form a well-supported clade in SSU rDNA as a separate family. phylogenies, supporting the erection of a separate Abe´ (1967a, p. 378) reported that ‘‘comparing family for the Phalacroma clade using either the Oxyphysis with Amphisolenia, one will notice that the name Oxyphysaceae, since Oxyphysis is a synonym of ventral area of Oxyphysis, covered largely by the pos- Phalacroma, or new family name, tentatively ‘‘Phala- terior sulcal plate, coincides with the same plate of cromaceae.’’ However, this change would need a Amphisolenia, not only in its shape but also in the clear definition of diagnostic characters for the structural relationships to the fission suture and the members of this putative separate family. In addi- longer right ventral hypothecal plate. This finding tion to difficulties in the classification into families suggests clearly a close phylogenetic affinity between and genera, even the classification into sectors is them.’’ Oxyphysis was thus placed in Amphisolenia- not supported by the molecular data. The members ceae (Loeblich 1982, Taylor 1987). Nevertheless, of Urophalacroma (P. mitra, P. rapa, and P. favus) are since Sournia (1984) established the family Oxy- morphologically similar species (Fig. 2, o–w), but, physaceae, the most extended practice in recent surprisingly, P. favus branched in a subclade differ- classifications has been to consider O. oxytoxoides as ent from P. mitra and P. rapa (Fig. 6). This exempli- the only member of its own family (Steidinger and fies the difficulties to find stable morphological Tangen 1997, Hastrup Jensen and Daugbjerg 2009). characters for the establishment of families, genera, However, Balech (1988, p. 201) observed that the or even sections within the dinophysoid dinoflagel- plate arrangement of Oxyphysis appeared to be close lates. to that of Dinophysis (he considered Phalacroma as a synonym) and placed it within the family Dinophys- This is a contribution to the project DIVERPLAN-MED sup- ported by a postdoctoral grant to F. G. of the Ministerio Espa- aceae. Our molecular data support Balech’s view. A n˜ol de Educacio´n y Ciencia #2007-0213. We acknowledge morphological intermediate between Phalacroma financial support from the French CNRS and the ANR Biodi- and Oxyphysis might be represented by a dinophy- versity program (ANR BDIV 07 004-02 ‘Aquaparadox’). We soid specimen with an elongate epitheca from the thank A. Nowaczyk for the BOUM samples and P. C. Silva for Gulf of Mexico illustrated by Balech (1967, pp. 92– his comments on nomenclature. 3, figs. 32–37). We provide the first micrograph of that nondescribed species from Lugol’s-fixed mate- Abe´, T. H. 1967a. The armored Dinoflagellata: II. Prorocentridae and Dinophysidae (A). Publ. Seto Mar. Biol. Lab. 14:369–89. rial collected the tropical western Pacific Ocean Abe´, T. H. 1967b. The armored Dinoflagellata: II. Prorocentridae (Fig. 4o). and Dinophysidae (B) – Dinophysis and its allied genera. Publ. O. oxytoxoides as well as other Phalacroma species, Seto Mar. Biol. Lab. 15:37–78. such as P. rotundatum, feed on by sucking Abe´, T. H. 1967c. The armored Dinoflagellata: II. Prorocentridae the prey cytoplasm by myzocytosis (Hansen 1991, and Dinophysidae (C) – Ornithocercus, Histioneis, Amphisolenia and others. Publ. Seto Mar. Biol. Lab. 15:79–116. Inoue et al. 1993). Oxyphysis seems to be specialized Balech, E. 1967. Dinoflagelados nuevos o interesantes del Golfo de on tintinnid preys with elongated loricae. Inoue Me´xico y Caribe. Rev. Mus. Argent. Cienc. Nat. Bernardino et al. (1993) observed that the distinctive elongated Rivadavia Hidrobiol. 2:77–126. epitheca of Oxyphysis intruded into the lorica of the Balech, E. 1977. Estructura de Amphisolenia bidentata Schro¨der (Dinoflagellata). Physis 37:25–32. preys. It can be speculated that Oxyphysis has Balech, E. 1980. On the thecal morphology of dinoflagellates with recently diverged morphologically to adapt to its species emphasis on cingular and sulcal plates. Anales Centro particular prey. The change in its general appear- Cienc. Mar. Limnol. Univ. Nac. Auton. Mex. 7:57–68. ance would be associated with a change in the dis- Balech, E. 1988. Los Dinoflagelados del Atla´ntico Sudoccidental. position of the plates when compared with the Publ. Espec. Inst. Esp. Oceanogr. 1:1–310. Edgar, R. C. 2004. MUSCLE: multiple sequence alignment with other Phalacroma species. The elongation of the cell high accuracy and high throughput. Nucleic Acids Res. 32: body would also imply a modification in the disposi- 1792–7. tion and shape of the thecal plates that has classi- Fensome, R. A., Taylor, F. J. R., Norris, G., Sarjeant, W. A. S., cally been used to support the placement in a Wharton, D. I. & Williams, G. L. 1993. A Classification of Living and Fossil Dinoflagellates. Am. Mus. Nat. Hist., Micropaleontol- separate family. ogy special publication number 7. Sheridan Press, Hanover. A morphological character commonly used for Pennsylvania, 351 pp. the generic separation in dinophysoid dinoflagel- Gaarder, K. 1954. Dinoflagellatae from the ‘‘Michael Sars’’ North lates is the height of the epitheca in Dinophysis and Atlantic Deep-Sea Expedition 1910. Rep. Sci. Res. ‘‘Michael Sars’’ Phalacroma (Kofoid and Skogsberg 1928, Hallegraeff N. Atl. Deep-Sea Exped. 2:1–62. Go´mez, F. 2003. Checklist of Mediterranean free-living dinoflagel- and Lucas 1988). Recently, Hastrup Jensen and lates. Bot. Mar. 46:215–42. Daugbjerg (2009) used this criterion for the delimi- Go´mez, F. 2005. A list of free-living dinoflagellate species in the tation of Phalacroma, which they restricted to cells world’s oceans. Acta Bot. Croat. 64:129–212. possessing large epithecae but <1 ⁄ 4 of the cell Go´mez, F. 2007. Synonymy and biogeography of the dinoflagellate genus Histioneis (Dinophysiales: Dinophyceae). Rev. Biol. Trop. length. The SSU and LSU rDNA phylogenies placed 55:459–77. Oxyphysis, with epithecae >1 ⁄ 4 of the cell length, in Go´mez, F., Lo´pez-Garcı´a, P. & Moreira, D. 2009. Molecular phy- the Phalacroma subclade closely related to the type logeny of the ocelloid-bearing dinoflagellates Erythropsidinium of Phalacroma. Consequently, the description of and Warnowia (Warnowiaceae, Dinophyceae). J. Eukaryot. Phalacroma should be amended. Most Phalacroma Microbiol. 56:440–5. 406 FERNANDO GO´ MEZ ET AL.

Hallegraeff, G. M. & Lucas, I. A. N. 1988. The marine dinoflagellate Akademische Verlagsgesellschaft, Leipzig, Germany, pp. 1– genus Dinophysis (Dinophyceae): photosynthetic, neritic and 617. non-photosynthetic, oceanic species. Phycologia 27:25–42. Sournia, A. 1984. Classification et nomenclature de divers dino- Handy, S. M., Bachvaroff, T. R., Timme, R. E., Coats, D. W., Kim, S. flagelle´s marins (Dinophyceae). Phycologia 23:345–55. & Delwiche, C. F. 2009. Phylogeny of four Dinophysiacean Steidinger, K. A. & Tangen, K. 1997. Dinoflagellates. In Tomas, C. genera (Dinophyceae, Dinophysiales) based on rDNA R. [Ed.] Identifying Marine Phytoplankton. Academic Press, San sequences from single cells and environmental samples. Diego, pp. 387–598. J. Phycol. 45:1163–74. Stein, F. R. 1883. Der Organismus der Infusionsthiere Nach Eigenen Hansen, P. J. 1991. Dinophysis – a planktonic dinoflagellate genus Forschungen in Systematischer Reihenfolge Bearbeitet. III. Abth. II. which can act both as a prey and a predator of a ciliate. Mar. Ha¨lfte. Die Naturgeschichte der Arhtodelen Flagellaten. Wilhelm Ecol. Prog. Ser. 69:201–4. Engelmann, Leipzig, Germany, 30 pp., 25 pls. Hastrup Jensen, M. & Daugbjerg, N. 2009. Molecular phylogeny of Tai, L. S. & Skogsberg, T. 1934. Studies on the Dinophysoidae, selected species of the order Dinophysiales (Dinophy- marine armored dinoflagellates, of Monterey Bay, California. ceae)—testing the hypothesis of a dinophysioid radiation. Arch. Protistenkde 82:380–482. J. Phycol. 45:1136–52. Taylor, F. J. R. 1976. Dinoflagellates from the International Indian Inoue, H., Fukuyo, Y. & Nimura, Y. 1993. Feeding behavior of Ocean Expedition. Bibl. Bot. 132:1–234. dinoflagellate, Oxyphysis oxytoxoides, on ciliates. Bull. Taylor, F. J. R. 1987. and classification. In Taylor, F. J. R. Soc. Jpn 40:9–17. [Ed.] The Biology of Dinoflagellates. Botanical Monographs 21. Jobb, G., von Haeseler, A. & Strimmer, K. 2004. TREEFINDER: a Blackwell, Oxford, UK, pp. 723–31. powerful graphical analysis environment for molecular phy- logenetics. BMC Evol. Biol. 4:18. Jørgensen, E. 1923. Mediterranean Dinophysiaceae. Rep. Dan. Oce- Supplementary material anogr. Exped. Mediterr. 2 (Biol.) J.2.:1–48. Kofoid, C. A. 1926. On Oxyphysis oxytoxoides gen. nov., sp. nov. The following supplementary material is avail- A dinophysoid Dinoflagellata convergent toward the peridi- able for this article: nioid type. Univ. Calif. Publ. Zool. 28:203–16. Kofoid, C. A. & Skogsberg, T. 1928. The Dinophysoideae. Mem. Appendix S1. A note on the spelling of the Mus. Comp. Zool. Harv. 51:1–766. Loeblich III, A. R. 1982. Dinophyceae. In Parker, S. P. [Ed.] Synopsis supergeneric names derived from Dinophysis and Classification of Living Organisms. McGraw-Hill, New York, authored by P. C. Silva. pp. 101–15. Lo´pez-Garcı´a, P., Rodrı´guez Valera, F., Pedro´s-Alio´, C. & Moreira, Figure S1. Maximum likelihood phylogenetic D. 2001. Unexpected diversity of small in deep-sea tree of Dinophysales LSU rDNA sequences, based Antarctic plankton. Nature 409:603–7. on 862 aligned positions. Nguyen, N. van, Omura, T., Furuya, K. & Fukuyo, Y. 2008. Dinophysis (Dinophyceae) in the pelagic waters of central and western Pacific. La Mer 46:29–36. Table S1. List of new SSU rDNA sequences of Norris, D. R. & Berner Jr, L. D. 1970. Thecal morphology of se- Dinophysales used for the phylogenetic analysis. lected species of Dinophysis (Dinoflagellata) from the Gulf of Accession numbers, geographic origin, and col- Mexico. Contrib. Mar. Sci. 15:145–92. lection dates are provided. Pavillard, J. 1916. Recherches sur les pe´ridiniens du Golfe de Lion. Trav. Inst. Bot. Univ. Montpellier Se´r. Mixte 4:9–70. This material is available as part of the online Pavillard, J. 1930. Sur quelques formes inte´ressantes ou nouvelles du phytoplankton (Diatome´es et Pe´ridiniens) des croisie`res article. du Prince Albert Ier de Monaco. Bull. Inst. Oceanogr. (Monaco) 558:1–12. Please note: Wiley-Blackwell are not responsi- Philippe, H. 1993. MUST, a computer package of management ble for the content or functionality of any supple- utilities for sequences and trees. Nucleic Acids Res. 21:5264–72. mentary materials supplied by the authors. Any Saitou, N. & Nei, M. 1987. The neighbor-joining method: a new queries (other than missing material) should be method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406–25. directed to the corresponding author for the Schiller, J. 1933. Dinoflagellatae (Peridineae). In Rabenhorst, L. article. [Ed.] Kryptogamen Flora von Deutschland. vol. 10, no. 3. I Teil.