phylogenetic position of attheya longicornis and attheya septentrionalis (bacillariophyta)

PHYLOGENETIC POSITION OF ATTHEYA LONGICORNIS AND ATTHEYASEPTENTRIONALIS (BACILLARIOPHYTA)1

Sebastiaan W. Rampen,2 Stefan Schouten, F. Elda Panoto, Maaike Brink

Department of Marine Organic Biogeochemistry, NIOZ Royal Netherlands Institute for Sea Research, PO Box 59,

1790 AB Den Burg, the Netherlands

Robert A. Andersen

Bigelow Laboratory for Ocean Sciences, PO Box 475, West Boothbay Harbor, Maine 04575, USA

Gerard Muyzer,3 Ben Abbas3

Department of Biological Oceanography, NIOZ Royal Netherlands Institute for Sea Research, PO Box 59,


and Jaap S. Sinninghe Damste

Department of Marine Organic Biogeochemistry, NIOZ Royal Netherlands Institute for Sea Research, PO Box 59,


The phylogenetic position of diatoms belonging tothe genus Attheya is presently under debate. Speciesbelonging to this genus have been placed in the sub-classes Chaetocerotophycidae and Biddulphiophyci-dae, but published phylogenetic trees based on 18SrDNA, morphology, and sexual reproduction indicatethat this group of diatoms may be a sister group ofthe pennates. To clarify the position of Attheya, westudied the morphology, 18S rDNA, 16S rDNA of thechloroplasts, the rbcL large subunit (LSU) sequencesof the chloroplasts, and the sterol composition ofthree different strains of Attheya septentrionalis(Østrup) R. M. Crawford and one strain of Attheyalongicornis R. M. Crawford et C. Gardner. These datawere compared with data from more than 100 otherdiatom species, covering the whole phylogenetic tree,with special emphasis on species belonging to thegenera that have been suggested to be related to thegenus Attheya. All data suggest that the investigated At-theya species form a separate group of diatoms, andthere is no indication that they belong to either theChaetocerotophycidae or the Biddulphiophycidae.Despite applying these various approaches, we wereunable to determine the exact phylogenetic positionof the investigated Attheya species within the diatoms.

Key index words: 16S rDNA; 18S rDNA; Attheya;Bacillariophyta; Biddulphiophycidae; Chaetocero-tophycidae; morphology; rbcL; sterols

Abbreviations: BI, Bayesian inference; BSTFA,N,O-bis(trimethylsilyl)trifluoroacetamine; GC, gas

chromatography; GC ⁄ MS, gas chromatography/massspectrometry; ML, maximum likelihood; NJ, neigh-bor joining; PP values, posterior probability values

Historically, diatoms have been separated into twomorphological categories, centrics and pennates,distinguished by the organizational pattern of thevalves (Schutt 1896). The genus Attheya falls in thegrey area that separates the archetypal centrics fromtheir pennate counterparts. Unlike many non-chain-forming diatoms, Attheya is usually seen anddescribed in girdle view. Its valves are elliptical, thereare extensive girdle bands, and two long hornsextend outward from each valve. Therefore, it super-ficially resembles single cells of the common genusChaetoceros. In the type description, Attheya decoraT. West was described being ‘‘precisely like Striatellaunipunctata in miniature’’ (West 1860, p. 152).According to West (1860, p. 153), A. decora appearedto ‘‘unite Striatella with Chaetoceros,’’ yet todayStriatella is treated as an araphid pennate diatom,and Chaetoceros as a centric diatom. However, West(1860) placed Attheya next to Chaetoceros based onsimilarities of the spiny processes. Since West’s publi-cation, diatoms related to A. decora have been trans-ferred among the genera Chaetoceros, Gonioceros, andAttheya (Evensen and Hasle 1975, Drebes 1977,Round et al. 1990, Crawford et al. 1994). For exam-ple, Chaetoceros septentrionalis Østrup was transferredto Gonioceros by Round et al. (1990), but thenCrawford et al. (1994) transferred it to Attheya. Evennow, the taxonomic position of A. septentrionalis andA. longicornis is uncertain because of significant dif-ferences in morphology and plastid type compared

1Received 30 July 2007. Accepted 13 October 2008.2Author for correspondence: e-mail [email protected] address: Department of Biotechnology, Delft University

of Technology, Julianalaan 67, 2628 BC Delft, the Netherlands.

J. Phycol. 45, 444–453 (2009)� 2009 Phycological Society of AmericaDOI: 10.1111/j.1529-8817.2009.00657.x

444

to the type species, A. decora. Specifically, Crawfordet al. (2000, p. 244) state, ‘‘two taxa stand out asbeing unlike the others – A. decora and A. armata.’’Based on A. decora, Attheya was placed in the subclassChaetocerotophycidae (family Attheyaceae) withChaetoceros, Gonioceros, and Bacteriastrum (family Chae-tocerotaceae) (Round et al. 1990). Subsequently,Crawford et al. (1994) placed Attheya in the subclassBiddulphiophycidae.

Although Attheya has been classified within thecentrics, remarkable similarities between Attheyaand pennate diatoms have been reported. Drebes(1977) and Chepurnov and Mann (2004) notedsimilarities between A. decora and the araphid pen-nates Rhabdonema and Striatella, particularly con-cerning the shape and stellate arrangement oftheir chloroplasts. They also discussed similaritiesin their sexual reproduction, especially the arrange-ment of auxospores and oogonia in Attheya com-pared to auxospores and gametangial thecae of theabove araphid pennates (see also Von Stosch1958). Rhabdonema and Striatella have elliptical tolanceolate valves and numerous girdle bands, sothat the body of the frustule resembles that ofAttheya. However, horns are absent in both genera,Rhabdonema forms epiphytic tabular colonies, andStriatella produces a mucilage stalk from an apicalpore plate. Furthermore, the annulus of Attheyaspecies is highly elongated and resembles the rib-like sternum typical of araphid pennate diatoms(Sims et al. 2006). Finally, Attheya species occur inbenthic habitats or sea ice. Generally, pennates aredistributed in benthic habitats, while centric dia-toms typically occur in the plankton. Althougheach of above features may occur in one or morecentric diatom genera, the combination of all thesepennate-like properties in Attheya suggests that thegenus Attheya may be derived from an immediateancestor of the pennates. This hypothesis is sup-ported by phylogenetic trees based on 18S rDNAby Sinninghe Damste et al. (2004) (Attheya specieswere incorrectly named Gonioceros), Kooistra et al.(2007), and Sorhannus (2007), although bootstrapsupport from the latter two studies is relatively low.

In an effort to clarify the phylogenetic position ofAttheya, we examined the morphology, 18S rDNA,16S rDNA, and RUBISCO LSU (rbcL) sequences,and the sterol composition of three strains ofA. septentrionalis and one of A. longicornis. This infor-mation was compared with that of more than 100other diatom species, representing the whole dia-tom phylogenetic diversity, with special emphasis onspecies belonging to genera that have been postu-lated as close relatives of Attheya, that is, diatoms ofthe Chaetocerotophycidae and Biddulphiophycidae.

MATERIALS AND METHODS

Cultivation. Typically, nonaxenic diatom cultures (Table S1in the supplementary material) were grown under optimal

conditions (see http://ccmp.bigelow.org, http://www.ccap.ac.uk, and http://www.marine.csiro.au/microalgae/index.html). The cultures were harvested for lipid and DNA analysesat the end of the log phase by filtration onto 0.7 lm glass fiberfilters. The filters were frozen directly and stored until furtheranalysis.

EM. For TEM, cells were gently pelleted, and the pellet wasresuspended in nitric acid to remove the organic material. Thecleaned frustules were washed several times with deionizedwater to remove all traces of acid. The frustules wereresuspended in water, placed on a carbon-coated pioloformgrid, and allowed to settle. The excess water was removed usinga pointed piece of filter paper. For SEM, frustules wereprepared by first removing some organic matter with bleachand then removing the remaining organic matter with 2Nsulfuric acid. The frustules were rinsed repeatedly with deion-ized water and dried onto aluminium foil attached to the SEMstub. The dried material was lightly coated with evaporatedcarbon (Denton Vacuum desk-model sputter-coater ⁄ evapora-tor, Desk IV, with carbon accessory; Denton Vacuum, Moores-town, NJ, USA) to avoid electrical charging. TEM was carriedout using a Zeiss 902A transmission electron microscope, whileSEM was carried out using a Zeiss Supra model 25 fieldemission scanning electron microscope (Carl Zeiss, Thorn-wood, NY, USA).

Molecular phylogeny. DNA extraction and purification havebeen described previously by Sinninghe Damste et al. (2004).Phylogenetic trees obtained from 18S rDNA sequences havepreviously been reported by Sinninghe Damste et al. (2004)and Rampen et al. (2007).

For this study, we performed PCR amplification andsequencing of the 16S rDNA and rbcL genes of the chloroplast.Amplification of the chloroplast 16S rDNA genes has beendescribed by Muyzer et al. (G. Muyzer, B. A. Abbas, S. W.Rampen, S. Schouten, and J. S. Sinninghe Damste, unpub-lished). The rbcL was amplified using two different primer setsgiving two overlapping fragments, A and B. Fragment A(1,100 bp) was made by primers NDrbcL2 [5¢-AAA AGT GACCGT TAT GAA TC-3¢ (Daugbjerg and Andersen 1997) andreverse primer 5¢-ATT TGD CCA CAG TGD ATA CCA-3¢(Corredor et al. 2004)]. Fragment B (800 bp) was made byprimer 5¢-GAT GAT GAR AAY ATT AAC TC-3¢ (Corredor et al.2004) and reverse primer 5¢-GTG TCT CAG CGA AAT CAG C-3¢(Fox and Sorhannus 2003). For both reactions, a mastermix of 10 mM dNTP’s, 0.5 lM of each primer, 1 unit of taqDNA polymerase, 10· diluted PCR-buffer, and 1 lL DNA(diluted in 10 mM Tris, 10·,100·, and 1000·), adjusted withwater to a 25 lL volume, was used. PCR conditions includedan initial denaturation step of 5 min at 94�C followed by 35cycles of 1 min denaturation at 94�C, 1 min primer anneal-ing at 52�C, 2 min primer extension followed by a finalextension of 7 min at 72�C. All outgroup sequences wereobtained from GenBank (http://www.ncbi.nlm.nih.gov/Genbank/) (Table S1).

The 18S rDNA sequences were generally �1,550–1,750nucleotides in length, 16S rDNA sequences were �1,000–1,450 nucleotides, and rbcL sequences were generally �1,000–1,350 nucleotides in length. For phylogenetic analyses, weselected 18S rDNA sequences longer than 1,500 nucleotides,16S rDNA sequences longer than 1,300 nucleotides, and rbcLsequences longer than 1,000 nucleotides. The 16S rDNA and18S rDNA sequences were aligned to sequences stored in theARB database; rbcL genes were aligned to each other. Highlyvariable regions of the DNA were excluded from calculationsusing a filter based on 50% base frequency across allsequences. Following this, 1,656 18S rDNA base positionsremained, of which 763 were variable; 1,428 base positionsof 16S rDNA remained, with 865 variable positions; and1,269 rbcL DNA base positions remained, with 658 variable

PLACEMENT OF ATTHEYA SPECIES 445

positions for phylogenetic analyses. Phylogenetic analyseswere performed using different algorithms (neighbor joining[NJ], maximum likelihood [ML], and Bayesian inference[BI]). NJ trees and ML trees were constructed using theprogram package PHYLIP, version 3.65 (Felsenstein andChurchill 1996). For NJ, 1,000 data sets were obtained fromthe original sequence sets using SEQBOOT, applied with thestandard settings. Distance matrices were obtained fromthese data sets using the F84 model incorporated inDNADIST. NJ trees, inferred from those distance matrices,were constructed using NEIGHBOR, implementing theclustering method described by Saitou and Nei (1987),using standard settings. Consensus trees were formed usingCONSENSE using standard settings. For 16S rDNA, anadditional consensus NJ tree was constructed from distancematrices obtained using the LogDet model, which is alsoincorporated in DNADIST. For rbcL, additional consensus NJtrees were constructed from distance matrices obtained froman rbcL data set consisting of only every first two codons,and from a protein data set, derived from the rbcL DNAdata set. Distance matrices from the data set consisting ofonly every first two codons were constructed in a similar wayas other distance matrices, using DNADIST; distance matricesfrom the protein data set were constructed using PROTDIST,applying the Jones-Taylor-Thornton matrix model. ML treeswere constructed from the sequence sets using DNAML(Felsenstein and Churchill 1996). The program ran with

standard settings, except that the option ‘‘Global rearrange-ments’’ was selected, and the option ‘‘Speedier but rougheranalysis’’ was turned off. BI consensus trees were constructedusing the program MrBayes 3.1.2. (Huelsenbeck et al. 2001),using the general-time-reversible (GTR) model with gamma-distributed rate variation across sites and a proportion ofinvariable sites, whereby three hot chains were run inaddition to the standard (cold) chain. For 18S rDNA, weran the Bayesian search during 1,000,000 generations, savingevery 1,000th tree, and the first 100 trees were discarded. For16S rDNA, we ran the Bayesian search during 2,000,000generations, saving every 100th tree, and the first 750 treeswere discarded. For rbcL, we ran the Bayesian search during6,400,000 generations, saving every 100th tree, and the first1,000 trees were discarded.

Sterol analyses. Filters were freeze dried and ultrasonicallyextracted as described by Schouten et al. (1998). An aliquotof the extracts was separated over alumina (Al2O3) usinghexane ⁄ dichloromethane (9:1, v ⁄ v) and dichloromethane ⁄methanol (1:1, v ⁄ v) to elute the apolar and sterol frac-tions, respectively. Prior to analyses by gas chromato-graphy (GC) and gas chromatography ⁄ mass spectrometry(GC ⁄ MS), sterol fractions were silylated by adding 25 lLBSTFA [N,O-bis(trimethylsilyl)trifluoroacetamine] and pyri-dine and heating the mixture at 60�C (20 min). GC andGC ⁄ MS analyses were performed as described by Schoutenet al. (1998).

Fig. 1. TEM whole mount and SEM images of Attheya. (a) TEM. A. septentrionalis showing the relatively short horns (arrows). CCMP2084. Scale bar, 3 lm. (b) TEM. Horn tip of A. septentrionalis showing terminal spines (arrow) around the open horn terminus. Scale bar,300 nm. CCMP2083. (c) TEM. Horn tip of A. septentrionalis showing terminal teeth and hoop-like siliceous bands (arrowheads).CCMP2084. Scale bar, 300 nm. (d) TEM. Horn of A. septentrionalis showing the three support rods (arrows). The hoop-like bands areattached to the outer two support rods. CCMP2084. Scale bar, 300 nm. (e) SEM. A. longicornis showing the horn tip. Note the bifurcatingpattern of the horn support rods (arrow) that terminate as spines (arrowhead). CCMP214. Scale bar, 210 nm. (f) TEM. Frustule parts ofA. longicornis showing the two very long horns that extend from either side of the valve (arrows). CCMP214. Scale bar, 2.5 lm. (g) SEM.A. longicornis showing the exterior of the valve (arrowhead), which has faint markings but no rimoportulae, and the underlying valvocopu-la. Note the valvocopula extension through the valve opening to provide short support for the base of the horn (arrow). CCMP214. Scalebar, 1 lm. (h) SEM. A. longicornis valvocopula and valve showing the inner surface of the valvocopula. CCMP214. Scale bar, 1 lm.

446 SEBASTIAAN W. RAMPEN ET AL.

RESULTS

Morphology. Examination of cultures CCMP2083and CCMP2084 using EM shows that they wereA. septentrionalis, while CCMP214 was identified asA. longicornis (Fig. 1, a–h). Like all species currentlyclassified as Attheya, these strains had four long horns,two extending out from each valve (Fig. 1, a, f–h).The horns consisted of numerous siliceous hoops orbands that were anchored to support rods extendingthe length of the horn (Fig. 1, b–d). At the distal endof the horn, the support rods bifurcated repeatedlyand ended as spines (Fig. 1, b, c, e). Numerous char-acters are used to distinguish species of Attheya (seeTable 1, Crawford et al. 1994, see also Crawford et al.2000 for the description of A. gausii). We identifiedA. septentrionalis based on its single cells (not chainforming), the lack of spines or rimoportulae on thevalves, the presence of one to two plastids, the pres-ence of relatively short horns, and the presence ofthree to four support rods in the horns. Crawfordet al. (1994) stated that A. septentrionalis has wavyhorns and four horn support rods; however, weobserved that the horns were wavy or straight and thatthe number of support rods was three. The numberof support rods was also three according to Evensenand Hasle (1975) and fig. 44 of Crawford et al.(1994), and therefore we consider the number ofsupport rods to be either three or four. A. longicorniswas very similar to A. septentrionalis; however, it waseasily distinguished by the very long horns (Fig. 1f).These two species differed significantly from the typespecies, A. decora, which has distinctly striated valves, arimoportula, copulae with well-defined pores, eightradially arranged chloroplasts, and the apparentabsence of horn support rods (Evensen and Hasle1975, Crawford et al. 1994).

Molecular phylogeny. Phylogenetic analyses wereperformed using different algorithms (NJ, ML, andBI). These algorithms generally resulted in treeswith similar clustering of species’ groups, thoughthere were some differences. We concentrate on theresults of the BI analyses of the different genes,after applying a 50% base frequency filter (seeFigs. 2–4), as this method has the advantage that itprovides posterior probability values (PP values),which are interpreted as the probability that a cladeis real (Huelsenbeck et al. 2001). We note the devia-tions from the BI trees of trees obtained using otheralgorithms. To visualize differences between thephylogenetic trees, taxonomic subdivisions based onmorphological characteristics and the 18S rDNAphylogeny (e.g., Kooistra et al. 2003) are coloreddifferently. Outgroups are colored black, radial cen-trics (except Thalassiosirales) brown, bi(multi) polarcentrics and Thalassiosirales (except Attheya) red,Attheya species pink, araphid pennates green, andraphid pennates blue. Furthermore, species belong-ing to the Biddulphiophycidae are labeled with a‘‘B,’’ and Chaetocerotophycidae are labeled with a‘‘C’’ as these have been postulated to be the closerelatives of Attheya species.

18S rDNA. Phylogenetic trees based on the 18SrDNA gene (e.g., Fig. 2) show one or more basalclades containing the radial centrics except the Tha-lassiosirales, designated as Clade 1 by Medlin andKaczmarska (2004), and all trees show a well-supported monophyletic clade containing the otherdiatoms, Clade 2 (Medlin and Kaczmarska 2004).All of the algorithms we used for reconstructing the18S rDNA phylogeny show at least one centric cladebranching off before the remaining diatoms ofClade 2 divide into a well-supported monophyletic

Table 1. Relative concentrations (as percentage of total sterol concentrations) of sterols in Attheya species, in speciesbelonging to the Chaetocerotophycidae and Biddulphiophycidae, and two species from the literature.

Species Origin

Sterolsa

C27 D5 C28 D5,22 C28 D5,24(28) C28 D5 C29 D5,24(28)E C29 D5,24(28)Z C29 D5

Attheya longicornis CCMP 214 23 44 5 5 21Attheya septentrionalis CCMP 2083 7 37 19 38Attheya septentrionalis CCMP 2084 6 33 17 1 9 33Attheya septentrionalis CS 425 ⁄ 03 29 36 10 24Bacteriastrum hyalinum CCMP 141 67 29 4Chaetoceros calcitrans NIOZ 12 78 2 6 <1Chaetoceros muelleri CCMP 1316 52 6 39 3Chaetoceros socialis NIOZ 92 8Chaetoceros sp. NIOZ 41 44 10 4Biddulphia sp. CCMP 147 32 37 7 <1Eucampia antarctica CCMP 1452 13 86Eucampia zoodiacus CCMP 386 87Odontella aurita CCMP 1108 14 52 9Odontella longicruris CCMP 1808 44 46 6LiteratureAttheya septentrionalis Ponomarenko et al. 2004 5 1 93Chaetoceros sp. Tsitsa-Tzardis et al. 1993 35 20 10 23

aC27 D5, cholest-5-en-3ß-ol; C28 D5,22, 24-methylcholesta-5,22-dien-3ß-ol; C28 D5,24(28), 24-methylcholesta-5,24(28)-dien-3ß-ol; C28

D5, 24-methylcholest-5-en-3ß-ol; C29 D5,24(28)E, 24-ethylcholesta-5,24(28)E-dien-3ß-ol; C29 D5,24(28)Z, 24-ethylcholesta-5,24(28)Z-dien-3ß-ol; C29 D5, 24-ethylcholest-5-en-3ß-ol.


pennate clade and a centric clade. Within thepennates, the raphid pennates always form a well-supported monophyletic clade. Importantly, in all ofthe phylogenetic trees, Attheya species grouptogether, with Neocalyptrella robusta as their sistergroup. In the NJ tree and the BI tree, this clade

forms the sister clade to the pennate diatoms, butthe relationship is not supported by high bootstrapor PP values (Fig. 2). In the ML tree, the Attheya-Neocalyptrella clade is the first of the bi(multi) polardiatom clades branching off after the Clade 1:Clade2 divergence, followed by a cluster containing

Fig. 3. Phylogeny of the diatoms, inferred with Bayesian inference analyses of 16S rDNA sequences of the chloroplast, after applyinga 50% base frequency filter. Blue = raphid pennates, green = araphid pennates, pink = Attheya species, red = bi(multi) polar centrics +Thalassiosirales, brown = radial centrics except Thalassiosirales, and black = outgroup species. The numbers at the nodes are posteriorprobability (PP) values (%). PP values below 50 are omitted. ‘‘B’’ indicates Biddulphiophycidae, and ‘‘C’’ indicates Chaetocerotophycidae.

Fig. 2. Phylogeny of the diatoms, inferred with Bayesian inference analyses of 18S rDNA, after applying a 50% base frequency-filter.Blue = raphid pennates, green = araphid pennates, pink = Attheya species, red = bi(multi) polar centrics + Thalassiosirales, brown = radialcentrics except Thalassiosirales, and black = outgroup species. The numbers at the nodes are posterior probability (PP) values (%). PP val-ues below 50 are omitted. ‘‘B’’ indicates Biddulphiophycidae, and ‘‘C’’ indicates Chaetocerotophycidae. The circles behind the sequencenames indicate presence or absence of 24-ethylcholest-5-en-3ß-ol in the analyzed cultures. Small open circles indicate no presence of24-ethylcholest-5-en-3ß-ol or concentrations below detection limit, small black circles indicate 24-ethylcholest-5-en-3ß-ol contributing to<10% of the total sterol composition, and large black circles indicate 24-ethylcholest-5-en-3ß-ol contributing to >10% of the total sterolcomposition. Species without circles behind their sequence name have not been analyzed for their sterol composition.


Chaetoceros species with Eucampia antarctica, andfinally, a clade of the remaining centric diatomsand a clade of pennate diatoms.

16S rDNA of the chloroplast. Phylogenetic treesobtained from 16S rDNA show more or less thesame pattern as observed for 18S rDNA: the radialcentrics are at the base of the tree, the bi(multi)polar diatoms together with Thalassiosirales andpennate diatoms form a monophyletic clade, andwithin this the pennates form a clade. However,none of 16S rDNA analyses placed the raphid

pennates in a monophyletic clade, and neither F84nor LogDet distance matrices resulted in high boot-strap values for Clade 2 or for the pennate diatomclade. Importantly, in all 16S rDNA trees, Attheyaspecies form a monophyletic clade, sister to the pen-nate diatoms, although not supported by high PP orbootstrap values.

rbcL sequences of the chloroplasts. In rbcL trees,related species generally cluster together, butthese clades only approximate to the expectedorder (Fig. 4). The presence of a radial centrics

Fig. 4. Phylogeny of the diatoms, inferred with Bayesian inference analyses of rbcL sequences, after applying a 50% base frequencyfilter. Blue = raphid pennates, green = araphid pennates, pink = Attheya species, red = bi(multi) polar centrics + Thalassiosirales, brown =radial centrics except Thalassiosirales, and black = outgroup species. The numbers at the nodes are posterior probability (PP) values (%).PP values below 50 are omitted. ‘‘B’’ indicates Biddulphiophycidae, and ‘‘C’’ indicates Chaetocerotophycidae.


clade, a bi(multi)polar clade including Thalassiosi-rales, a clade consisting of pennate diatoms, anda clade consisting of raphid pennates can be dis-cerned in the BI tree, but in NJ and ML trees,clusters of related species seem to be randomlyplaced. In rbcL trees, Bolidomonas does not fallas an outgroup but clusters within the centricdiatoms.

In the BI rbcL tree, A. septentrionalis groups withthree araphid pennate species, Fragilaria pinnata,Nanofrusulum shiloi, and Synedra fragilaroides, and inthe NJ and ML trees, Attheya forms the sister cladeto the Thalassiosirales. The low PP and bootstrapvalues for the subclass positions indicate low sup-port for the rbcL trees. This was not improved byanalyzing data sets with the third codon excluded,or rbcL protein data sets.

Sterols. The main sterols in the four Attheyaspecies are 24-methylcholesta-5,24(28)-dien-3ß-ol,24-methylcholest-5-en-3ß-ol, and 24-ethylcholest-5-en-3ß-ol (Table 1). The first two are also found asmajor sterols in the Chaetocerotophycidae and Bid-dulphiophycidae, but 24-ethylcholest-5-en-3ß-ol wasonly present in trace amounts or below detectionlimit (Table 1). 24-Methylcholesta-5,24(28)-dien-3ß-ol is probably the most common sterol in diatoms.It was present in all bi(multi) polar centric diatomsand Thalassiosirales, >50% of our araphid pennateand radial centric diatoms, and in >20% of all pen-nate diatoms we analyzed. 24-Methylcholest-5-en-3ß-ol was present in >50% of all radial, bi(multi) polar,araphid, and raphid pennate diatoms we analyzed;24-ethylcholest-5-en-3ß-ol was present in >50% of allradial centrics and raphid pennates, �45% of allaraphid pennate diatoms, and �35% of all bi(multi)polar centric diatoms we analyzed (Fig. 2). Accord-ing to our data set, no sterol is present in all pen-nate diatoms or in all araphid pennate diatoms.The most common sterol in araphid pennate dia-toms, 24-methylcholesta-5,22-dien-3ß-ol (present inalmost 70% of the araphid pennate diatoms), wasnot detected in Attheya.

DISCUSSION

As mentioned in the introduction, the classifica-tion of A. septentrionalis and A. longicornis haschanged over time. Several Attheya species haverecently been described using EM, and it is possi-ble that additional species will be described. Weare uncertain about the placement of A. septentrio-nalis and A. longicornis within the genus Attheyabecause of the morphological features that distin-guish them (and others) from the type species(see also Crawford et al. 2000). Molecular phyloge-netic analysis that includes the type species wouldbe valuable for resolving the classification ofA. septentrionalis and A. longicornis. Regardless ofthe generic assignment, these species may occupyan important phylogenetic position with regard to

the evolution of pennate diatoms. In a number ofstudies, similarities in habitat, frustule shape,annulus shape, shape and stellate arrangement ofthe chloroplasts, and sexual reproduction betweenspecies of Attheya and the pennate diatoms werereported, particularly between Attheya and Rhabdo-nema and Striatella (Von Stosch 1958, Drebes 1977,Chepurnov and Mann 2004, Sims et al. 2006).Phylogenetic trees inferred from 18S rDNA(Sinninghe Damste et al. 2004, Sorhannus 2007,Kooistra et al. 2007) place Attheya as a sister cladeto the pennate diatoms, albeit with low bootstrapsupport. The present study did not provideenough data to constrain unambiguously the phy-logenetic position of the Attheya. In most of ourtrees (NJ and BI from 18S rDNA, and NJ, BI, andML from 16S rDNA), Attheya species form a sistergroup of the pennate diatoms, but never withhigh bootstrap or PP support. Moreover, in the18S rDNA ML tree, Attheya forms the root ofClade 2, consisting of bi(multi) polar centrics,Thalassiosirales, and the pennate diatoms. The BIrbcL tree places Attheya in a clade consisting ofpennate species, F. pinnata, N. shiloi, and S. fragil-aroides, and in NJ and ML rbcL trees, it forms thesister clade of the Thalassiosirales. In all 18SrDNA trees, the clade containing Attheya also con-tains Neocalyptrella robusta; unfortunately, there areno 16S rDNA or rbcL sequences of N. robusta toconfirm this placement. This relationship betweenN. robusta and Attheya is not resolved in othertrees (Sinninghe Damste et al. 2004 as Rhizosoleniarobusta; Kooistra et al. 2007 as Calyptrella; orSorhannus 2007).

Although our results do not provide certaintyabout the exact phylogenetic position of Attheya,there is no indication that it belongs to either theChaetocerotophycidae or the Biddulphiophycidae.Placement of Attheya in these subclasses is also notsupported by their sterol composition. None ofthese diatom groups possesses unique sterols, butthe sterol composition of Attheya is unique, con-sisting of high concentrations of 24-methylcholes-ta-5,24(28)-dien-3ß-ol, 24-methylcholest-5-en-3ß-ol,and 24-ethylcholest-5-en-3ß-ol. Other diatom sterolstudies (Orcutt and Patterson 1975, Ballantineet al. 1979, Yamaguchi et al. 1986, Gladu et al.1991, Barrett et al. 1995, Veron et al. 1996) haveshown that besides Attheya, only a few diatomsfrom the Thalassiosirales contain similar sterol pat-terns consisting of high amounts of 24-methylcho-lesta-5,24(28)-dien-3ß-ol, 24-methylcholest-5-en-3ß-ol,and 24-ethylcholest-5-en-3ß-ol. Strong similaritiesbetween the sterol composition of our Attheya spe-cies and a species of Chaetoceros (B-13) reportedby Tsitsa-Tzardis et al. (1993) may indicate thatthe latter diatom was actually an Attheya species.Ponomarenko et al. (2004) reported the sterolcomposition of A. ussurensis, in which 24-ethylcho-lest-5-en-3ß-ol comprised 93% of all sterols. Unlike


our Attheya species, 24-methylcholesta-5,24(28)-dien-3ß-ol was absent, and 24-methylcholest-5-en-3ß-ol was only a minor sterol in A. ussurensis.High relative concentrations of 24-ethylcholest-5-en-3ß-ol in A. longicornis, A. septentrionalis, andA. ussurensis may indicate a relationship betweenthe three species. A. ussurensis most closely resem-bles A. decora (Stonik et al. 2006). As no molecu-lar data are available for these species, the exactrelationship between A. ussurensis, A. longicornis,and A. septentrionalis remains unresolved. Unfortu-nately, the sterol data do not provide informationabout the phylogenetic position of Attheya withrespect to the pennate diatoms.

CONCLUSIONS

Morphology, phylogenetic trees, and the sterolcomposition reported in this study indicate thatA. longicornis and A. septentrionalis form a separategroup of diatoms, and there is no indication thatthey belong to the Chaetocerotophycidae or theBiddulphiophycidae. None of the results questionsthe position of Attheya within the centric diatoms,but the results of this study provide no definiteanswer regarding their putative sister relationshipwith the pennate diatoms. Studies involving the typespecies of Attheya and Gonioceros should help resolvethe systematic relationships of A. longicornis andA. septentrionalis.

This work was supported by the Dutch Technology Founda-tion (STW) grant BAR-5275, by Grant 853.00.020 fromthe ALW coupled Biosphere–Geosphere programme of theNetherlands Organisation for Scientific Research (NWO),and by U.S. National Science Foundation grants 0444418 and0629564. We thank Dr. Wetherbee and two anonymousreviewers for their comments, which significantly improvedthe quality of this paper.

Ballantine, J. A., Lavis, A. & Morris, R. J. 1979. Sterols of thephytoplankton. Effects of illumination and growth stage.Phytochemistry 18:1459–66.

Barrett, S. M., Volkman, J. K. & Dunstan, G. A. 1995. Sterols of 14species of marine diatoms (Bacillariophyta). J. Phycol. 31:360–9.

Chepurnov, V. A. & Mann, D. G. 2004. Auxosporulation of Licmo-phora communis (Bacillarlophyta) and a review of mating sys-tems and sexual reproduction in araphid pennate diatoms.Phycol. Res. 52:1–12.

Corredor, J. E., Wawrik, B., Paul, J. H., Tran, H., Kerkhof, L.,Lopez, J. M., Dieppa, A. & Cardenas, O. 2004. Geochemicalrate-RNA integrated study: ribose-1,5-bisphosphate carboxyl-ase ⁄ oxygenase gene transcription and photosyntheticcapacity of planktonic photoautotrophs. Appl. Environ.Microbiol. 70:5459–68.

Crawford, R. M., Gardner, C. & Medlin, L. K. 1994. The genus Attheya.I. A description of four new taxa, and the transfer of Goniocerosseptentrionalis and G. armatas. Diatom Res. 9:27–51.

Crawford, R. M., Hinz, F. & Koschinski, P. 2000. The combinationof Chaetoceros gaussii (Bacillariophyta) with Attheya. Phycologia39:238–44.

Daugbjerg, N. & Andersen, R. A. 1997. Phylogenetic analyses of therbcL sequences from haptophytes and heterokont algae suggesttheir chloroplasts are unrelated. Mol. Biol. Evol. 14:1242–51.

Drebes, G. 1977. Cell structure, cell division and sexual reproduc-tion of Attheya docora West (Bacillariophyceae, Biddulphii-neae). Nova Hedwigia Beih. 54:167–78.

Evensen, D. L. & Hasle, G. R. 1975. The morphology of someChaetoceros (Bacillariophyceae) species as seen in the electronmicroscopes. Nova Hedwigia Beih. 53:153–84.

Felsenstein, J. & Churchill, G. A. 1996. A hidden Markov modelapproach to variation among sites in rate of evolution. Mol.Biol. Evol. 13:93–104.

Fox, M. G. & Sorhannus, U. M. 2003. RpoA: a useful gene for phylo-genetic analysis in diatoms. J. Eukaryot. Microbiol. 50:471–5.

Gladu, P. K., Patterson, G. W., Wikfors, G. H., Chitwood, D. J. &Lusby, W. R. 1991. Sterols of some diatoms. Phytochemistry30:2301–3.

Huelsenbeck, J. P., Ronquist, F., Nielsen, R. & Bollback, J. P. 2001.Bayesian inference of phylogeny and its impact on evolution-ary biology. Science 294:2310–4.

Kooistra, W. H. C. F., De Stefano, M., Mann, D. G. & Medlin, L. K.2003. The phylogeny of the diatoms. Prog. Mol. Subcell. Biol.33:59–97.

Kooistra, W. H. C. F., Gersonde, R., Medlin, L. K. & Mann, D. G.2007. The origin and evolution of the diatoms: their adapta-tion to a planktonic existence. In Falkowski, P. G. & Knoll, A.H. [Eds.] Evolution of Primary Producers in the Sea. ElsevierAcademic Press, Singapore, pp. 207–49.

Medlin, L. K. & Kaczmarska, I. 2004. Evolution of the diatoms: V.Morphological and cytological support for the major cladesand a taxonomic revision. Phycologia 43:245–70.

Orcutt, D. M. & Patterson, G. W. 1975. Sterol, fatty acid and ele-mental composition of diatoms grown in chemically definedmedia. Comp. Biochem. Physiol. 50B:579–83.

Ponomarenko, L. P., Stonik, I. V., Aizdaicher, N. A., Orlova, T. Y.,Popovskaya, G. I., Pomazkina, G. V. & Stonik, V. A. 2004.Sterols of marine microalgae Pyramimonas cf. cordata (Pra-sinophyta), Attheya ussurensis sp. nov. (Bacillariophyta) and aspring diatom bloom from Lake Baikal. Comp. Biochem. Physiol.B 138:65–70.

Rampen, S. W., Schouten, S., Abbas, B., Panoto, F. E., Muyzer, G.,Campbell, C. N., Fehling, J. & Sinninghe Damste, J. S. 2007.On the origin of 24-norcholestanes and their use as age-diag-nostic biomarkers. Geology 35:419–22.

Round, F. E., Crawford, R. M. & Mann, D. G. 1990. The Diatoms.Biology and Morphology of the Genera. Cambridge Univ. Press,Cambridge, UK, 747 pp.

Saitou, N. & Nei, M. 1987. The neighbor-joining method – a newmethod for reconstructing phylogenetic trees. Mol. Biol. Evol.4:406–25.

Schouten, S., Klein Breteler, W. C. M., Blokker, P., Schogt, N.,Rijpstra, W. I. C., Grice, K., Baas, M. & Sinninghe Damste, J. S.1998. Biosynthetic effects on the stable carbon isotopiccomposition of algal lipids: implications for deciphering thecarbon isotopic biomarker record. Geochim. Cosmochim. Acta62:1397–406.

Schutt, F. 1896. Bacillariales. In Engler, A. & Prantl, K. [Eds.] DieNaturlichen Planzenfamilien, Vol. 1. Wilhelm Engelmann,Leipzig, Germany, pp. 31–153.

Sims, P. A., Mann, D. G. & Medlin, L. K. 2006. Evolution of thediatoms: insights from fossil, biological and molecular data.Phycologia 45:361–402.

Sinninghe Damste, J. S., Muyzer, G., Abbas, B., Rampen, S. W.,Masse, G., Allard, W. G., Belt, S. T., et al. 2004. The rise of therhizosolenid diatoms. Science 304:584–7.

Sorhannus, U. 2007. A nuclear-encoded small-subunit ribosomalRNA timescale for diatom evolution. Mar. Micropaleontol. 65:1–12.

Stonik, I. V., Orlova, T. Y. & Crawford, R. M. 2006. Attheya ussurensissp. nov. (Bacillariophyta) – a new marine diatom from thecoastal waters of the Sea of Japan and a reappraisal of thegenus. Phycologia 45:141–7.

Tsitsa-Tzardis, E., Patterson, G. W., Wikfors, G. H., Gladu, P. K. &Harrison, D. 1993. Sterols of Chaetoceros and Skeletonema. Lipids28:465–7.


Veron, B., Billard, C., Dauguet, J.-C. & Hartmann, M.-A. 1996.Sterol composition of Phaeodactylum tricornutum as influencedby growth temperature and light spectral quality. Lipids31:989–94.

Von Stosch, H. A. 1958. Kann die oogame Araphidee Rhabdonemaadriaticum als Bindeglied zwischen den beiden grossenDiatomeengruppen angesehen werden? Ber. Dt. Bot. Ges.71:241–9.

West, T. 1860. Remarks on some Diatomaceae, new or imperfectlydescribed, and a new desmid. Trans. Microsc. Soc. Lond. 8:147–53.

Yamaguchi, T., Ito, K. & Hata, M. 1986. Studies on thesterols in some marine phytoplanktons. Tohoku J. Agr. Res.37:5–14.

Supplementary Material

The following supplementary material is avail-able for this article:

Table S1. Diatom and outgroup species, theirorigin and GenBank accession numbers.

This material is available as part of the onlinearticle.

Please note: Wiley-Blackwell are not responsiblefor the content or functionality of any supplemen-tary materials supplied by the authors. Any queries(other than missing material) should be directedto the corresponding author for the article.


phylogenetic position of attheya longicornis and attheya septentrionalis (bacillariophyta)

Documents