Proc. Natl. Acad. Sci. USA Vol. 94, pp. 11411–11416, October 1997 Evolution

Phylogenetic relationships between the Acantharea and the Polycystinea: A molecular perspective on Haeckel’s Radiolaria

LINDA AMARAL ZETTLER*†,MITCHELL L. SOGIN*, AND DAVID A. CARON‡

The Josephine Bay Paul Center in Comparative Molecular Biology and Evolution, Marine Biological Laboratory, 7 MBL Street, Woods Hole, MA 02543; and ‡Biology Department, Woods Hole Oceanographic Institution, 324 Redfield Building, Woods Hole, MA 02543

Communicated by Andrew H. Knoll, Harvard University, Cambridge, MA, August 11, 1997 (received for review May 9, 1997)

ABSTRACT Polycystine radiolaria are among few pro- mitochondria, golgi, and other major cellular machinery from tistan groups that possess a comprehensive fossil record the extracapsular region that contains the axopodial network available for study by micropaleontologists. The Polycystinea of the cell. Phaeodarian and polycystine radiolaria possess and the Acantharea, whose skeletons do not become fossilized, pores in their central capsules whereas acantharia do not. were once members of the class ‘‘Radiolaria’’ (‘‘Radiolaria’’ Despite these differences, the common use of radial symmetry sensu lato: Polycystinea, Phaeodarea, and Acantharea) origi- in cell body plan and shell architecture often results in a similar nally proposed by Haeckel but are now included in the general appearance among members of the Acantharea and superclass Actinopoda. Phylogenetic relationships within this the polycystine order Spumellarida. superclass remain largely enigmatic. We investigated the Today, the term ‘‘radiolaria’’ is generally an informal taxo- evolutionary relationship of the Acantharea and the Poly- nomic descriptor for members of the Polycystinea (Spumel- cystinea to other protists using phylogenetic analyses of larida and Nassellarida) and the Phaeodarea, even though 16S-like ribosomal RNA (rRNA) coding regions. We circum- Ernst Haeckel (5) used the term to include members of the vented the need to culture these organisms by collecting and class Acantharea. Of the 4417 species of organisms described maintaining reproductive stages that contain many copies of from collections of the Challenger Expedition, 3508 were new their genomic DNA. This strategy facilitated extraction of genomic DNA and its purification from symbiont and prey species of Radiolaria (5, 6). In his classification of the ‘‘Ra- DNA. Phylogenetic trees inferred from comparisons of 16S- diolaria,’’ Haeckel included four legions: the Acantharia, the like coding regions do not support a shared history between Spumellaria, the Nassellaria, and the Phaeodaria. This classi- the Acantharea and the Polycystinea. However, the monophyly fication was later modified (7–10) to exclude the Acantharia of the Acantharea and the separate monophyly of the Poly- from the Spumellaria, Nassellaria, and Phaeodaria. Despite cystinea (Spumellarida) are well supported by our molecular- later taxonomic revision, many of Haeckel’s original descrip- based trees. The acantharian lineage branches among crown tions of the Challenger Radiolaria persist today. Modern organisms whereas the polycystine lineage diverges before the systematists place acantharia in a class distinct from polycys- radiation of the crown groups. We conclude that the Actino- tines and phaeodaria but generally agree that these classes are poda does not represent a monophyletic evolutionary assem- members of the Actinopoda (2, 3, 11–13). blage and recommend that this taxonomic designation be The presence of axopodia and a central capsule are the discarded. predominant shared characters of the Acantharea and the Polycystinea. Most similarities occur between the Acantharea The superclass Actinopoda (1) includes protists with special- and the Polycystinea belonging to the order Spumellarida. ized microtubule-stiffened pseudopodia called ‘‘axopodia.’’ Most acantharia are polynucleated, but the presence of a single Current taxonomic schemes of the superclass Actinopoda nucleus in the acantharian genus Haliommatidium is a feature include the Acantharea, Polycystinea, and Phaeodarea and shared with the polycystines, which usually have only one most members of the Heliozoea (except the pedinellids and nucleus. Similarities also occur between the ‘‘gelatinous pel- heliomonads) (2, 3). licle’’ of some acantharia and that of certain Sphaerellarina, a One of the major distinctions between the Acantharea and suborder in the Spumellarida (14, 15). Both classes use SrSO4. the Polycystinea is the composition, architecture, and symme- Members of the Acantharea build entire skeletons of SrSO4 try of the skeleton, when present. Polycystines have a rich fossil whereas adult vegetative colonial spumellaria and swarmer record due to the presence of skeletal material comprised of cells of reproductive spumellaria contain individual crystals of silica, which is secreted by many members of this class and SrSO4 (2, 16–18). which becomes preserved in marine sediments. Polycystine The molecular diversity of the superclass Actinopoda is skeletons exhibit a range of morphologies from simple spicules unexplored. We sequenced the 16S-like rRNA genes of rep- to more elaborate latticed shells possessing radial spines. All resentatives of the Acantharea and Polycystinea to test the acantharia form skeletons comprised of monocrystals of stron- hypothesis that the Acantharea and the Polycystinea share a tium sulfate (SrSO ), which come together at the center of the 4 common evolutionary history. According to phenotypic sim- cell in a symmetrical fashion known as Mu¨ller’s Law (4). ilarities, the taxa chosen for this study represent the most Acantharian skeletons, however, are susceptible to dissolution closely related classes in the Actinopoda. Therefore, this study and are therefore absent from the fossil record. In all phae- odaria, polycystines, and members of the acantharian order addresses the appropriateness of the higher taxon designation Arthracanthida, the central capsule or capsular wall divides the Actinopoda and rekindles the debate over the best definition cell into an intracapsular region that includes the nucleus, of ‘‘Radiolaria.’’

The publication costs of this article were defrayed in part by page charge Abbreviations: rRNA, ribosomal RNA; SrSO4, strontium sulfate; FITC, fluorescein isothiocyanate. payment. This article must therefore be hereby marked ‘‘advertisement’’ in Data deposition: The sequence in this paper have been deposited in accordance with 18 U.S.C. §1734 solely to indicate this fact. the GenBank database (accession nos. AF018158–AF018163). © 1997 by The National Academy of Sciences 0027-8424͞97͞9411411-6$2.00͞0 †To whom reprint requests should be addressed. e-mail: amaral@ PNAS is available online at http:͞͞www.pnas.org. evol5.mbl.edu.

11411 Downloaded by guest on September 25, 2021 11412 Evolution: Amaral Zettler et al. Proc. Natl. Acad. Sci. USA 94 (1997)

MATERIALS AND METHODS hybridizations on acantharia were carried out as described (26) using biotinylated probes and detection with fluorescein iso- Sample Collection. Individual specimens were collected by thiocyanate (FITC)–avidin solution (20 ␮g͞ml in 100 mM divers using glass or polycarbonate jars. Specimens were NaHCO3-buffered saline, pH 8.2). Acantharia were fixed for maintained in 0.22 ␮m of filtered Sargasso Sea water in glass 1 h at 4°C in 1X Histochoice fixative (Amresco, Euclid, OH) culture tubes with brine shrimp (Artemia salina) as food. All diluted in 0.22 ␮m of filtered Sargasso Sea water. Individuals acantharia were collected off of the southwestern coast of were then transferred to gel-subbed slides, overlaid with 0.05% Bermuda in September 1994. Acantharian samples included agarose, and allowed to dry overnight. Probe was added to a Haliommatidium sp. (BBSR 235: Order: Symphyacanthida, final concentration of 5 ng͞␮l. Probe experiments consisted of Family: Pseudolithidae) and Chaunacanthid 218 (BBSR 218: a negative control (incubation in fluorescein-labeled avidin Order: Chaunacanthida). Polycystine radiolarian specimens with no probe added), a positive control (biotinylated EUK included one solitary and three colonial forms, all from the 1209Rbio added), and two separate acantharian-specific probe order Spumellarida. The solitary spumellarian Thalassicolla experiments using A497bio and A899bio oligonucleotides, nucleata (BBS 3: Family: Thalassicollidae) was collected in respectively. Hybridizations were carried out at 42°C for 6–8 May 1992, colonial spumellarian Collosphaera globularis- h and washed at 45°C. Cells were mounted in Citifluor huxleyi (BBSR 173: Family: Collosphaeridae) in May 1994, and immersion oil (Citifluor, Kent, U.K.) and viewed on a Zeiss colonial spumellaria Sphaerozoum punctatum (CR4: Family: Axiophot equipped for epifluorescence microscopy. Sphaerozoidae) and Collozoum serpentinum (CR16: Family: We used colorimetry-based in situ hybridizations for the Sphaerozoidae) in May 1995. Specimens used for in situ colonial spumellarian samples to avoid severe autofluores- hybridizations were collected in September and October of cence. Hybridizations were carried out on colonial spumellar- 1995. ian samples using the GIBCO͞BRL In Situ Hybridization and DNA Extraction, Amplification, Cloning, and Sequencing. Detection System with the following modifications for use with A natural amplification of DNA occurs before swarmer cell rRNA and larger sarcodines. Colonies were preserved in 1X release from reproductive acantharia and spumellaria. Many Histochoice with 95% ethanol added in a ratio of 4:1 for 1 h species of acantharia and spumellaria either consume or expel at 4°C, transferred to 70% ethanol, and held overnight at 4°C. endocytoplasmic symbiotic algae just before swarmer cell Aliquots of preserved central capsules from a single colony formation. We harvested specimens at this point of the life were placed on silanated glass slides (Midwest Scientific; St. cycle to enrich for sarcodine DNA and to reduce the potential Louis, MO) and allowed to air dry. Slides were then baked at of amplifying nontarget DNA. 65°C for 1 hour to remove endogenous alkaline phosphatase Individual central capsules or cysts were rinsed in 0.22 ␮m activity. Hybridizations were carried out in 50 ␮l-capacity of filtered sea water and transferred to lysis or PCR buffer Probe-Clip ‘‘Press-to-Seal’’ incubation chambers and holders solution. Central capsules from T. nucleata specimens were (Midwest Scientific, St. Louis). Four probe experiments were lysed and genomic DNA was extracted and precipitated ac- carried out using central capsules from the same colony: a cording to standard protocols (19). Central capsules or indi- negative control incubation (streptavidin-alkaline phospha- vidual acantharian cysts from taxa other than T. nucleata were tase conjugate with no probe added), a negative probe control rinsed as above, placed in modified 1X PCR buffer [50 mM (A899bio acantharian probe added), a positive probe control KCl͞10 mM Tris, pH 8.3͞2 mM MgCl2͞0.001% gelatin͞1.0% (EUK 502bio and EUK 1209bio added), and a colonial Nonidet P-40 (Sigma)], and lysed at 95°C for 10 minutes. spumellarian probe experiment (R906bio and R1451bio add- Thalassicolla nucleata, C. globularis-huxleyi, and Haliomma- ed). The final concentration of probes in all experiments was tidium sp. 16S-like rDNAs were PCR-amplified using eukary- 1ng͞␮l. otic primers for nuclear 16S-like rRNA coding regions (20). Hybridizations were conducted according to the manufac- Templates were cloned into either M13 or the plasmid vector turer’s instructions for ‘‘DNA Detection’’ with the above pCRII (Invitrogen). Colonial spumellarian or acantharian modifications and the omission of any steps specifically re- specific probes described below were used as primers in quired for DNA targets. Slides were hybridized for 8 h, and combination with Medlin primers to amplify the remaining probe detection was carried out according to manufacturer’s samples that were then sequenced directly. Double-stranded protocol with levamisole (Sigma) added at 200 ␮g͞ml upon sequencing of the rRNA coding regions was accomplished addition of alkaline phosphatase conjugate to remove endog- using either Sequenase Ver. 2.0 (United States Biochemical) enous alkaline phosphatase activity. Developed slides were and existing rDNA primers (20, 21) or using a Li-Cor 4000L permanently mounted in Crystal͞Mount (Biomedia, Foster automated sequencer and Sequitherm cycle sequencing pro- City, CA) and observed on a Zeiss standard microscope tocols developed by Li-Cor (Lincoln, NE) (22). equipped with phase microscopy. In Situ Hybridizations. In situ whole cell hybridizations using Phylogenetic Analysis. The 16S-like rRNA sequences of oligonucleotide probes complementary to the presumptive acantharian and spumellarian samples were aligned against sarcodine 16S-like rRNA verified that the sequence data were a subset of sequences representing a range of eukaryotic derived from the sarcodine DNA and not from contaminating evolutionary lineages maintained by the Ribosomal Data- algal symbionts or prey material. We designed probes that base Project (23). Initially, we aligned our sarcodine se- would specifically hybridize with colonial radiolaria or acan- quences with the subset of eukaryotic 16S-like rRNA se- tharia but no other eukaryotes in the Ribosomal Database quences in the Ribosomal Database Project according to Project (23). The biotin-labeled probes designed for acanthar- conservation of primary structure. The alignment was re- ian samples were A497bio, 5Ј-TCATTCCAATCAACT- fined by juxtaposing regions that define evolutionarily con- CAC-3Ј and A899bio, 5Ј-TCGTCATACAAAGGTCCA-3Ј. served secondary structures. Phylogenetic analyses were The colonial spumellarian probes were R906bio, 5Ј- based on 1369 positions considered to be in unambiguous AACGATAAAATACTAATA-3Ј and R1451bio, 5Ј-TATTG- alignment. Structural similarities were converted to evolu- TAGCCCGTGCGCT-3Ј. Eukaryote-specific, 5Ј-biotinylated tionary distances using the Jukes Cantor (27) correction, and probes that hybridize with all known eukaryotic 16S-like a phylogeny was inferred by the method of Olsen (28). One rRNAs served as positive controls. These were EUK502Rbio; hundred bootstrap replicates were conducted, and a consen- 5Ј-ACCAGACTTGCCCTCC-3Ј (24) and EUK1209Rbio; 5Ј- sus tree was obtained using PHYLIP Ver. 3.5 (29). Phyloge- GGGCATCACAGACCTG-3Ј (25). netic trees were also inferred by the maximum likelihood In situ hybridizations were carried out using both fluores- method using the program FASTDNAML (30) using a gener- cence and colorimetric detection methods. Fluorescence in situ alized two-parameter model of evolution (31) and the max- Downloaded by guest on September 25, 2021 Evolution: Amaral Zettler et al. Proc. Natl. Acad. Sci. USA 94 (1997) 11413

imum parsimony method using the program PAUP, Ver. 3.1.1 DISCUSSION (32). The maximum parsimony tree was obtained from a The relative positions of the acantharia and the spumellaria in consensus of 100 bootstrap replications that were conducted molecular phylogenetic trees indicate that the presence of using a heuristic search option with random addition se- axopodia, a capsule membrane, and the ability to metabolize quence, 10 replicates, and the tree bisection–reconnection SrSO4 should be reconsidered as reliable phylogenetic markers algorithm. Identical phylogenetic analyses were performed for these taxonomic groups. Our molecular study of acanthar- with Phreatamoeba balamuthi removed from the data set to ian and spumellarian phylogeny confirms that axopodia have determine stability of the relative branching order of the evolved more than once and most likely represent convergent acantharia and the polycystine radiolaria. structures created in response to similar ecological constraints through evolutionary time (34–38). An independent evolution RESULTS of axopodia within the chromistan Pedinellea of the Heliozoea already has been suggested on morphological grounds (2). In situ hybridization experiments confirmed the origin of the Given the results of this study, retention of the superclass acantharian and spumellarian sequences (Fig. 1). Acantharian- Actinopoda seems inappropriate, as does the adoption of the specific probes specifically hybridized to the acantharia (Fig. 1, new phylum Radiozoa, which has been described as a modern- F and H) but not to colonial spumellaria (Fig. 1J). Likewise, day Radiolaria sensu lato (2, 3). colonial spumellarian probes specifically hybridized with co- The presence of a central capsule and the ability to secrete lonial spumellaria (Fig. 1L). SrSO4 are two synapomorphies for the Radiozoa (2). However, The%GϩC content of the 16S-like rRNA gene for the central capsule found in spumellaria and that which exists Haliommatidium sp. and Chaunacanthid 218 was 44% and in one order of the Acantharea (the Arthracanthida) differ (15, 45%, respectively, which was similar to many of the other taxa 37). The acantharian central capsule is nonperforated and of used in the analyses. However, spumellarian % G ϩ C content a different cellular origin than that found in polycystines and values (35–38%) were similar to that of Entamoeba gingivalis phaeodaria. Furthermore, the presence of central capsules in (34%) and were low relative to typical eukaryotic values that the Arthracanthida, an order that is considered more derived are usually close to 50%. The lengths of 16S-like rRNA coding than other orders of acantharia that lack central capsules regions in base pairs for the spumellaria ranged from 1770 to (37–39), suggests that ‘‘central capsules’’ may have evolved 1798, which are typical of other eukaryotes. more than once. The phylogenetic trees inferred by the distance matrix, The occurrence of SrSO4 in both acantharia and spumellaria is another feature cited as evidence of their common ancestry. maximum parsimony (Fig. 2), and maximum likelihood (data SrSO constitutes the main skeletal material in acantharia. not shown) methods clearly rejected a common ancestry 4 Vegetative adults of colonial spumellaria, however, also are between the acantharia and the spumellaria. Numbers at nodes known to contain crystals of SrSO4 in their central capsules, represent bootstrap values as a percentage of 100 resamplings and the biflagellated swarmers of all spumellaria examined of the data set. Only bootstrap values Ͼ50% (which we thus far contain crystals of SrSO4 in membrane-bound vesicles. interpret as relative measures of confidence for specific topo- However, metabolism of SrSO4 is not unique to acantharia and logical features of the tree) are shown. Both the distance and spumellaria. Crystals of SrSO4 have been observed in the parsimony trees placed the spumellaria branching as a diverg- desmid alga Closterium littorale (40), in Chara, the ‘‘stonewort’’ ing lineage below the ‘‘crown’’ group (33), those taxa repre- freshwater plant (40), and in loxodid ciliates (41). The role of senting major eukaryotic assemblages simultaneously radiat- SrSO4 in Chara and the loxodid ciliates appears to be gravi- ing from the node labeled with a bootstrap value of 85͞65. perception (40, 41). A similar function has been proposed in Both methods revealed a poorly resolved branching point for the desmid algae (40). This function apparently has never been the acantharia (Haliommatidium sp. and Chaunacanthid 218) proposed for the membrane-bound crystals found in spumel- among the crown radiation. The relative positions of the larian swarmers. Instead, Anderson (18) proposed that SrSO4 acantharia and the polycystine radiolaria were not affected by crystals may serve a function in buoyancy control but noted removal of Phreatamoeba balamuthi (data not shown). Re- that silica or calcium compounds, which occur at higher moval of Phreatamoeba balamuthi from the data set resulted in concentrations in sea water, would be better candidates for this higher bootstrap support values for the node leading to the purpose. Anderson also suggested that strontium may be of crown (89͞98) with acantharia branching among crown groups some physiological importance to the spumellaria but did not and spumellaria branching before the crown groups. elaborate on what this requirement might be. One possibility A low bootstrap value of 67% was obtained for the branch- is that strontium serves a similar function in spumellaria as in ing of the spumellaria with Entamoeba gingivalis in the parsi- some gastropods where it is required for proper shell devel- mony analysis, but this support was not observed in the opment (42). One inconsistency of the potential importance of distance analysis nor the topology of the maximum likelihood strontium in spumellarian skeletal development is that spu- analysis (data not shown). No other potential immediate mellarian species that lack skeletal material (e.g., T. nucleata and Collozoum spp.) have swarmers with crystalline SrSO common ancestors for the spumellaria were indicated by these 4 inclusions. Also, the lack of SrSO crystals in acantharian data. The position of the spumellaria relative to other groups 4 swarmer cells is inconsistent with the hypothesis that SrSO4 branching below the crown varied between the distance and serves a similar function in both the Acantharea and the parsimony analyses. Therefore, the exact branching order of spumellarian polycystines. the spumellaria also remains unresolved at this time. The The absence of SrSO4 crystals in swarmer cells of the monophyly of the acantharia and the separate monophyly of Acantharea implies another difference between SrSO4 utili- the spumellaria, however, were well supported (100% in all zation in acantharia and spumellaria. Acantharian swarmer cases). cells lack SrSO4 crystals, yet their reproductive life cycle stage The branching patterns within the spumellaria in both occurs deeper in the euphotic zone (37). This suggests a need distance and parsimony analyses showed the solitary spumel- for buoyancy control. Perhaps acantharian cells use myonemes larian T. nucleata branches before the colonial spumellaria. and other cellular structures as hydrostatic devices to regulate Although the bootstrap support for this node was only 52% in depth in the water column. In addition, many species of the parsimony analysis, a higher bootstrap value (77%) was acantharia form cysts also comprised of SrSO4 that aid in the obtained with distance methods. sinking of swarmer cells to depth. The greater density of SrSO4 Downloaded by guest on September 25, 2021 11414 Evolution: Amaral Zettler et al. Proc. Natl. Acad. Sci. USA 94 (1997)

FIG.1. In situ hybridization of Histochoice-preserved specimens using oligodeoxynucleotide probes complementary to the 16S-like (small subunit) rRNA sequences of acantharia (A–H) and colonial spumellaria (I–L). For both acantharian and colonial spumellarian cells, probes conjugated to biotin were detected by either FITC–avidin or streptavidin–alkaline phosphatase-conjugated secondary labels. For the acantharian cells, hybridization was detected by epifluorescence microscopy with settings specific for FITC excitation (B, D, F, and H). Colonial spumellarian cells were viewed using phase contrast microscopy, and hybridizations were detected colorimetrically using the localized, purple precipitate of the enzymatic reaction of alkaline phosphatase on nitroblue tetrazolium and 5-bromo-4-chloro-3-indolylphosphate substrates. (A–H) Four different acantharian cells of the same species with corresponding transmitted light and epifluorescence photomicrographs of the same cell. Epifluorescence photomicrographs were taken with an integral camera system using a FITC filter set consisting of a 450- to 490-nm band-pass excitation filter, a 510-nm long-pass dichroic mirror, and a 515- to 565-nm band-pass emission filter. Fuji 100 ASA Provia color slide film was used for fluorescence pictures. Transmitted light photomicrographs of samples were taken with an Olympus OM4-T camera using Kodak 160 speed Tungsten film. All exposure times for a set of samples (i.e., negative control, positive control, taxon-specific probes) were kept constant so that the relative intensity was indicative of probe binding. (Bars ϭ 75 ␮m for A–H.) (A and B) The negative control to which only FITC–avidin was added. Note the minimal background fluorescence of the cell under epifluorescence (B). (C and D) The positive control treatment to which a eukaryotic-specific probe designed to target all eukaryotes (EUK 1209R) was added. (E–H) Probing with two different acantharian probes (F, A497; H, A899). (I–L) Hybridization results for single individuals within the same spumellarian colony. (I) The negative control to which only the streptavidin-alkaline phosphatase-conjugated secondary label was added. (J) A negative probe control to which an acantharian probe was added. (K) The results of the positive control hybridization with eukaryote probes (EUK502R and EUK1209R). (L) Hybridization with colonial spumellarian probes (R906 and R1451). (Bars ϭ 35 ␮m for I–L.)

relative to silica may also explain why polycystine radiolaria, that an extant relative shares some of the unique features which can possess siliceous skeletons, use the heavier SrSO4 in possessed by the Acantharea. We conclude that the ability to their swarmers. Of interest, phaeodarian radiolaria, which live deposit SrSO4 is a less reliable phylogenetic marker than the deeper in the water column, have lost the capacity to deposit ability to construct elaborate skeletal architecture using SrSO4. SrSO4 and, like acantharia, lack SrSO4-containing swarmer For example, the closest relative to the acantharia may also cells. Whatever function served, the presence of SrSO4 in these possess strontium-containing skeletons. Among the possibili- marine protists and its singular utilization as the structural ties is Podactinelius sessilis (43), first described on the Deut- compound in the skeletons of acantharia warrant further schen Su¨dpolar Expedition of 1901–1903. This genus possesses scrutiny in the evolution of this group as do their unique spines of SrSO4 (400–500); however, these spines are not non-actin-containing myonemes. arranged in the characteristic geometric pattern observed in all Which protistan lineage shares the most recent ancestry with members of the Acantharea. The genus was once included as the Acantharea? The branching pattern of the acantharia was a separate order Actineliida in the class Acantharea; however, shallow relative to that of the spumellaria, so it is likely that it has been relegated recently to an uncertain taxonomic acantharia diversified more recently than the spumellaria and affinity (11). Downloaded by guest on September 25, 2021 Evolution: Amaral Zettler et al. Proc. Natl. Acad. Sci. USA 94 (1997) 11415

FIG. 2. The inferred phylogeny for the acantharia and the spumellaria. A distance tree is shown. Bootstrap values for distance analyses are given above the line whereas maximum parsimony values are below. A dash indicates that the bootstrap value for that node was below 50% in the method used for phylogeny reconstruction. The bar insert corresponds to 10 changes per 100 nt positions.

The determination of the nearest relative of the Spumel- tinued use of this definition in anything but a historical larida remains uncertain. It may be difficult to determine the perspective is unjustified in view of the results described phenotype of the most recent common ancestor of the Spu- herein. mellarida if the long branches in the spumellarian lineage are interpreted as evidence of a very ancient divergence. Although We thank Drs. Annika Sanfilippo and William Riedel for reprints the fossil record of spicule-bearing forms (Sphaerozoidae) and translated Russian references, Dr. Rebecca Gast for critical extends to the Lower Oligocene (44) and that of the Col- reading of this manuscript, Dr. Alexi Shalapyonok for assistance with translating Russian manuscripts, and Gaspar Taroncher–Oldenburg losphaeridae to the base of the Miocene (45), even more for assistance in German translation. We also thank Drs. Anthony ancient origins are possible in view of the existence of extant Michaels, O. Roger Anderson, and Neil Swanberg for help with skeletonless forms that would not be preserved in the sedi- identifications of specimens and helpful discussions. Drs. Anthony ments. As an alternative hypotheses, the long branch lengths of Michaels and Hank Trapido-Rosenthal opened their labs at the the spumellaria may be explained as the result of a rapidly Bermuda Biological Station for Research. Frances Howes and numer- evolving lineage. ous other volunteers assisted with blue water diving collection of It is assumed that the Nassellarida that represent the second specimens. This report is based upon work supported under a National order included in the Polycystinea are closely related to the Science Foundation Graduate Fellowship. Further financial support is acknowledged in part from the National Science Foundation (OCE- Spumellarida (46). The molecular phylogenetic position of the 9314533) (to D.A.C.) and from the Scurlock Fund for research at the Phaeodarea is also unknown and deserves consideration. As Bermuda Biological Station for Research, the Grant-In-Aid Fund of for Haeckel’s Radiolaria and the definition of the Radiolaria the Bermuda Biological Station for Research, funds from the Woods sensu lato (Polycystinea, Phaeodarea, and Acantharea), con- Hole Coastal Research Center, the Education Office of the Woods Downloaded by guest on September 25, 2021 11416 Evolution: Amaral Zettler et al. Proc. Natl. Acad. Sci. USA 94 (1997)

Hole Oceanographic Institution, and the Lerner Gray Fund for Marine 22. Middendorf, L. R., Bruce, J. C., Bruce, R. C., Eckles, R. D., Research of the American Museum of Natural History (to L.A.A.Z.), Grone, D. L., Roemer, S. C., Slonicker, G. D., Steffens, D. L., National Institutes of Health (GM32964), and the Unger G. Vettlesen Sutter, S. L., Brumbaugh, J. A. & Patonay, G. (1992) Electro- Foundation (to M.L.S.). This is Woods Hole Oceanographic Institu- phoresis 13, 487–494. tion Contribution No. 9547. 23. Olsen, G. J., Overbeek, R., Larsen, N., Marsh, T. L., McCaughey, M. J., Maciukenas, M. A., Kuan, W. -M., Macke, T. J., Xing, Y. 1. Calkins, G. N. (1909) Protozoo¨logy (Lea & Febiger, New York), & Woese, C. R. (1992) Nucleic Acids Res. 20, Suppl., 2199–2200. pp. 1–349. 24. Amann, R. I., Binder, B. J., Olson, S. W., Devereux, R. & Stahl, 2. Cavalier-Smith, T. (1993) Microbiol. Rev. 57, 953–994. D. A. (1990) Appl. Environ. Microbiol. 56, 1919–1925. 3. Corliss, J. O. (1994) Acta Protozool. 33, 1–51. 25. Giovannoni, S. J., DeLong, E. F., Olsen, G. J. & Pace, N. R. 4. Mu¨ller, J. (1858) in Abhandlungen der Ko¨niglische Akademie der (1988) J. Bacteriol. 170, 720–726. Wissenschaften zu Berlin, ed. Dummler, F. (Deutsche Akademie 26. Lim, E. L., Amaral, L. A., Caron, D. A. & DeLong, E. F. (1993) der Wissenschaften zu Berlin, Berlin), pp. 1–62. Appl. Environ. Microbiol. 59, 1647–1655. 5. Haeckel, E. (1887) in The Voyage of the H. M. S. Challenger, ed. 27. Jukes, T. H. & Cantor, C. R. (1969) in Mammalian Protein Thompson, C. W. & Murray, J. (Her Majesty’s Stationary Office, Metabolism, ed. Munro, H. N. (Academic, New York), pp. 21–32. London), pp. 1–1760. 28. Olsen, G. J. (1988) Methods Enzymol. 164, 793–812. 6. Anderson, O. R. (1983) Radiolaria (Springer, New York), pp. 29. Felsenstein, J. (1985) Evolution 39, 783–791. 1–355. 30. Olsen, G. J., Matsuda, H., Hagstrom, R. & Overbeek, R. (1994) 10, 7. Deflandre, G. (1952) in Traite´de Pale´ontologie, ed. Piveteau, J. Comput. Appl. Biosci. 41–48. J. Mol. Evol. 29, (Masson, Paris), pp. 303–313. 31. Kishino, H. & Hasegawa, M. (1989) 170–179. 32. Swofford, D. L. (1991) PAUP: Phylogenetic Analysis Using Par- 8. Deflandre, G. (1953) in Traite´de Zoologie, ed. Grasse´, P. P. simony (Illinois Natural History Survey, Champaign, IL). (Masson, Paris), pp. 389–436. 33. Knoll, A. H. (1992) Science 256, 622–627. 9. Tregouboff, G. (1953) in Traite´de Zoologie, ed. Grasse´, P. P. 34. Merinfeld, E. G. (1978) Biosystems 10, 29–30. (Masson, Paris), pp. 271–320. 35. Shulman, S. S. & Reshetnyak, V. V. (1980) Realno li cush- 10. Goll, R. M. & Merinfeld, E. G. (1979) in The Encyclopedia of chestvovanie v cisteme prosteishikh nadklassa Actinopoda? (About Paleontology , ed. Fairbridge, R. W. & Jablonski, D. (Dowden, the Possibility of the Superclass Actinopoda Existence in the System Hutchinson, & Ross, Stroudsburg), pp. 673–684. of Protozoa?), ed. Skarlato, O. A. (Academy of Science of the 11. Levine, N. D., Corliss, J. O., Cox, F. E. G., Deroux, G., Grain, J., USSR, Proceedings of the Zoological Institute, Leningrad, Honigberg, B. M., Leedale, G. F., Loeblich, A. R. I., Lom, J., USSR), pp. 23–41. Lynn, D., Merinfeld, E. G., Page, F. C., Poljansky, G., Sprague, 36. Merinfeld, E. G. (1981) Phylogenetic Relationships Among the V., Vavra, J. & Wallace, F. G. (1980) J. Protozool. 27, 37–58. Actinopoda (Warszawa, Poland), p. 244. 12. Lee, J. J., Hutner, S. H. & Bovee, E. C. (1985) An Illustrated Guide 37. Reshetnyak, V. V. (1981a) in Acantharia (Acanthariea, Protozoa) to the Protozoa (Society of Protozoologists, Lawrence, KS), pp. from the World Ocean (Nauka, SSSR, Leningrad), p. 224. 1–629. 38. Hollande, A., Cachon, J. & Cachon-Enjumet, M. (1965) Protis- 13. Febvre, J. (1990) in Handbook of Protoctista, ed. Margulis, L., tologica 1, 91–112. Corliss, J. O., Melkonian, M. & Chapman, D. J. (Jones and 39. Strelkov, A. A. & Reshetnyak, V. V. (1974) in Chemoradiology of Bartlett, Boston), pp. 363–379. the Pelagical and Benthal (Metals and Their Radionuclei in Hy- 14. Hollande, A. & Enjumet, M. (1960) Arch. Museum Paris VII, drobionts and in Their Environment Medium), ed. Polikarpov, 1–134. G. G. (Naukova Dumka, Kiev), pp. 95–191. 15. Massera Bottazzi, E. (1978) Boll. Zool. 45, 133–144. 40. Raven, J. A., Smith, F. A. & Walker, N. A. (1986) in Biominer- 16. Dogel, V. A. (1950) Zoologicheski Zhurnal (Zoological Journal) alization in Lower Plants and Animals, ed. Leadbeater, B. S. C. & 29, 562–565. Riding, R. (Clarendon, New York), pp. 125–139. 17. Hollande, A. & Martoja, R. (1974) Protistologica X, 603–609. 41. Fenchel, T. & Finlay, B. J. (1986) J. Protozool. 33, 69–79. 18. Anderson, O. R. (1981) in Silicon and Siliceous Structures in 42. Bidwell, J. P., Paige, J. A. & Kuzirian, A. M. (1986) Biol. Bull. 170, Biological Systems, eds. Simpson, T. L. & Volcani, B. E. (Springer, 75–90. New York), pp. 347–379. 43. Schro¨der, O. (1907) Deutsche Su¨dpolar-Expedition 1901–1903 9, 19. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1984) Molecular 227–236. Cloning: A Laboratory Manual, (Cold Spring Harbor Lab. Press, 44. Bjørklund, K. R. & Goll, R. M. (1979) J. Paleontol. 53, 1293–1326. Plainview, NY). 45. Riedel, W. R. (1967) in The Fossil Record, ed. Harland, W. B. 20. Medlin, L., Elwood, H. J., Stickel, S. & Sogin, M. L. (1988) Gene (Geol. Soc. London, London), 291–298. 71, 491–499. 46. Cachon, J., Cachon, M. & Estep, K. W. (1990) in Handbook of 21. Elwood, H. J., Olsen, G. J. & Sogin, M. L. (1985) Mol. Biol. Evol. Protoctista, ed. Margulis, L., Corliss, J. O., Melkonian, M. & 2, 399–410. Chapman, D. J. (Jones and Bartlett, Boston), pp. 334–346. Downloaded by guest on September 25, 2021