ORIGINAL ARTICLE

Absence of co-phylogeny indicates repeated diatom capturein dinophytes hosting a tertiary endosymbiont

Anže Žerdoner Čalasan1& Juliane Kretschmann1

& Marc Gottschling1

Received: 28 July 2017 /Accepted: 26 October 2017 /Published online: 18 November 2017# Gesellschaft für Biologische Systematik 2017

Abstract Tertiary endosymbiosis is proven throughdinophytes, some of which (i.e. Kryptoperidiniaceae) haveengulfed diatom algae containing a secondary plastid.Chloroplasts are usually inherited together permanently withthe host cell, leading to co-phylogeny. We compiled a diatomsequence data matrix of two nuclear and two chloroplast loci.Almost all endosymbionts of Kryptoperidiniaceae found theirclosest relatives in free-living diatoms and not in otherharboured algae, rejecting co-phylogeny and indicating thatresident diatoms were taken up by dinophytes multiple timesindependently. Almost intact ultrastructure and insignificantgenome reduction are supportive for young, if not recentevents of diatom capture. With their selective specificity onthe one hand and extraordinary degree of endosymbiotic flex-ibility on the other hand, dinophytes hosting diatoms sharemore traits with lichens or facultatively phototrophic ciliatesthan with green algae and land plants. Time estimates indicatethe dinophyte lineages as consistently older than the hosteddiatom lineages, thus also favouring a repeated uptake of en-dosymbionts. The complex ecological role of dinophytesemploying a variety of organismic interactions may explaintheir high potential and plasticity in acquiring a great diversityof plastids.

Keywords Chloroplast . Dinoflagellates . Dinotoms .

Endosymbiosis . Evolution .Mutualism

Introduction

As one of the important groups of unicellular eukaryotic lifeforms, Dinophyceae (or Dinoflagellata under zoological no-menclature) are nothing less but diverse from every singlepoint of view (Morden and Sherwood 2002; Pochon et al.2012; Gottschling and McLean 2013; Burki 2014). Togetherwith Ciliata and Apicomplexa (= Sporozoa), Dinophyceaebelong to Alveolata and are a well-supported monophyleticgroup based on molecular data and numerous unique anatom-ical characteristics (Harper et al. 2005; Medlin and Fensome2013; Keeling et al. 2014; Janouškovec et al. 2017). Comparedto all other eukaryotes, the genome of dinophytes is highlyunusual with respect to structure and regulation (Moreno Díazde la Espina et al. 2005; Wisecaver and Hackett 2011). Thenucleus contains chromosomes that are permanently condensedthroughout cell development (displaying a liquid crystallinestate: Rill et al. 1989) except in the DNA replication stage(Dodge 1966; Rizzo 2003). Dinophyte morphology also ex-hibits unique traits such as the coiled transverse flagellum asso-ciated with a transverse groove termed the ‘cingulum’ (Taylor1980; Fensome et al. 1999; Leander and Keeling 2004;Okamoto and Keeling 2014).

Despite the fact that they represent only about a half of alldinophyte species (Taylor et al. 2008), photosynthetically ac-tive taxa have gained chloroplasts in numerous advancedways (Gagat et al. 2014; Dorrell and Howe 2015), excludingthe most basic primary engulfment (primary endosymbiosis).The most common and widespread chloroplast type found indinophytes indicates a secondary engulfment event (second-ary endosymbiosis) of a red algae (Dodge 1989; Jeffrey 1989;

Electronic supplementary material The online version of this article(https://doi.org/10.1007/s13127-017-0348-0) contains supplementarymaterial, which is available to authorized users.

* Marc Gottschlinggottschling@bio.lmu.de

1 Department Biologie, Systematische Botanik und Mykologie,GeoBio-Center, Ludwig-Maximilians-Universität München,Menzinger Str. 67,, 80 638 Munich, Germany

Org Divers Evol (2018) 18:29–38https://doi.org/10.1007/s13127-017-0348-0

Janouškovec et al. 2017). Some dinophyte predators captureadditional plastids of their prey and keep them up to severalweeks in an evolutionary phenomenon known askleptoplastidy (Schnepf and Elbrächter 1999; Stoecker 1999;Gast et al. 2007; Takano et al. 2014). Other dinophytes havetaken up a different endosymbiont in a more permanent wayand have brought the endosymbiosis even further to a tertiarylevel by engulfing organisms such as diatoms, which alreadypossessed secondarily gained chloroplasts (Keeling 2004,2010; Hehenberger et al. 2014).

An example for a chloroplast replacement at a second-ary level of endosymbiosis is Lepidodinium M.Watan.,S.Suda, I.Inouye et al. (Gymnodiniaceae s.str.) enclosinga green alga (Kamikawa et al. 2015). Tertiary endosym-biosis, primarily found in dinophytes, is even more com-plex from the symbiont’s point of view. Tertiary plastidsare present in, for example, species of Dinophysis Ehrenb.(Dinophysaceae) possessing cryptomonad-derived sym-bionts separated from the host by two membranes(Schnepf and Elbrächter 1988; Garcia-Cuetos et al.2010). Taxa such as Karenia Gert Hansen & Moestrupand Karlodinium J.Larsen (Brachidiniaceae) possess chlo-roplasts that have originated from haptophyte algae(Hansen et al. 2003; Gast et al. 2007). Endosymbiontreduction is severe in this group, and the three-membrane layer as part of the chloroplast is the onlystructure left of the engulfed organism. An ancestralperidinin chloroplast may have co-existed with the newlygained chloroplast of haptophyte origin but has been losteventually (Nosenko et al. 2006; Patron et al. 2006;Takano et al. 2014). However, there is no ultrastructuralevidence that Brachidiniaceae have ever possessed twodifferent functional chloroplasts.

Based on ultrastructural analyses of those dinophytesharbouring diatoms (Kryptoperidiniaceae, commonly referredto as ‘dinotoms’: Imanian et al. 2010), the endosymbiontkeeps not only its photosynthesis machinery but also the func-tional nucleus with protein encoding genes, a substantialamount of cytosol and mitochondria (McEwan and Keeling2004; Imanian and Keeling 2007; Imanian et al. 2012; Takanoet al. 2008). However, numerous studies also show a loss ofmotility, cell wall and mitotic division (Dodge 1971; Tomasand Cox 1973). Despite this major autonomy and minimalreduction, the resident diatom is present in the host cell in allstages of development, which indicates synchrony of life his-tory between both host and endosymbiont (Figueroa et al.2009). Moreover, diatom mitochondria seem to be almost un-affected by endosymbiosis, retaining nearly all of their char-acteristics and functions (Imanian and Keeling 2007; Imanianet al. 2012). Engulfed diatoms due to tertiary endosymbiosisreplace the ancestral peridinin chloroplast, whose remnantsare considered to be a unique form of an eyespot (Dodge1983; Horiguchi et al. 1999; Moestrup and Daugbjerg 2007;

Takano et al. 2008) surrounded by a triple-membrane structure(Horiguchi and Pienaar 1994).

All Kryptoperidiniaceae constitute a monophyletic groupas a part of the Peridiniales (Pienaar et al. 2007; Takano et al.2008; Gottschling and McLean 2013; Janouškovec et al.2017; Price and Bhattacharya 2017; Kretschmann et al. inpress), but their endosymbionts originate from at least threedifferent diatom lineages. The currently accepted scenario de-scribes the single acquisition of a Nitzschia Hassall-relatedorganism (present today in Durinskia Carty & El.R.Cox,Galeidinium Tam. & T.Horig. and KryptoperidiniumEr.Lindem.: Tamura et al. 2005; Pienaar et al. 2007; Zhanget al. 2011a) that has been replaced by either Cyclotella(Kütz.) Bréb.-related (in freshwater UnruhdiniumGottschling: Takano et al. 2008; Zhang et al. 2011b; Youet al. 2015) or by Chaetoceros Ehrenb.-related symbionts (inBlixaea Gottschling: Horiguchi and Takano 2006) in two fur-ther evolutionary steps. The phenomenon of monophyletichosts and polyphyletic symbionts has thus been explainedby ‘serial replacement’ (Horiguchi and Takano 2006;Takano et al. 2008) of an already engulfed diatom. However,this idea has been developed based on single-locus analyseswith a rather limited diatom taxon sample.

While preparing a similar study of our own, we took inter-est in the study of Yamada et al. (2017). The study advocatesserial replacement, considering a few diatom acquisitionevents, after which endosymbionts are established, maintainedand inherited. Subsequently, endosymbionts are in permanentassociation with the host cell and should thus find their closestrelatives among other endosymbionts. Co-phylogenetic struc-tures should therefore be present between separately derivedtress of diatoms and their hosting dinophytes. This hypothesisis not explicitly phrased in Yamada et al. (2017), although it isthe necessary conceptual basis to understand the origin ofdinophytes harbouring tertiary endosymbionts and the dy-namics of this (endo)symbiont/host association. Our data pro-vide evidence for a repeated acquisition of endosymbioticdiatoms through their evolutionary history and for the rejec-tion of overall co-phylogeny. Tertiary endosymbiosis presentin Kryptoperidiniaceae may have happened recently in com-parison to many other algal groups, in which severe reductionhas taken place leaving behind only the chloroplasts and theirmembranes.

Materials and methods

From the dinophyte endosymbionts of Durinskia oculata(F.Stein) Gert Hansen & Flaim and Kryptoperidinium sp.,we sequenced the nucleomorph small subunit (SSU) andInternal Transcribed Spacers (ITS) and plastid rbcL loci aspreviously described (Kretschmann et al. in press). In orderto place dinophyte endosymbionts on diatom reference trees, we

30 Žerdoner Čalasan A. et al.

gathered extensive sequence data (Tab. S1, mainly provided byAlverson et al. 2007; Lee et al. 2013; Li et al. 2015; Medlin2015; Theriot et al. 2015). For a global diatom analysis, wecompiled all diatom operational taxonomic units (OTUs)possessing at least two nuclear (SSU and large subunit: LSU)and two chloroplast loci (rbcL and either psbA or psbC), alongwith sequences from the Kryptoperidiniaceae possessing at leastone nuclear (SSU) and one chloroplast (rbcL) locus. In a con-siderable number of cases, further plastid loci such as atpB,psaA, psaB and psbA were added, when they were availablefrom the same strains. For two additional analyses at taxonom-ically subordinate levels (i .e. Bacillariaceae andStephanodiscaceae), we performed BLAST searches (Altschulet al. 1997) using the endosymbiont sequences as query to ob-tain all available sequences of closely related diatoms. Thealignments included then all diatom and endosymbiont se-quences (excluding ITS due to the high divergence), regardlessof the locus coverage.

Separate matrices were constructed, aligned using‘MAFFT’ v6.502a (Katoh and Standley 2013) andconcatenated afterwards with SeaView v4.6.1 (Gouy et al.2010). Phylogenetic analyses were carried out using both aML (Stamatakis 2014) as well as a Bayesian approach(Ronquist et al. 2012), as described in detail previously(Gottschling et al. 2012). The aligned matrices are availableas *.nex files upon request. All new DNA sequences weredeposited in GenBank.

I n o r d e r t o e s t i m a t e d i v e r g e n c e t im e s i nKryptoperidiniaceae, we constituted a concatenated alignment(defined by the nuclear rRNA loci SSU, ITS and LSU plusmitochondrial MT-CO1 and MT-CYB) that was shown to besuitable for phylogenetic inference (Gottschling et al. 2012;Gu et al. 2013; Tillmann et al. 2014). It included allPeridiniales, of which sequence information of all three nucle-ar regions was available, along with a representative set ofrRNA sequences covering the known diversity ofKryptoperidiniaceae (irrespectively whether they were com-plete or not). The phylogeny was dated using BEAST v1.8.3(Drummond et al. 2012), with settings recommended for in-terspecific data that might or might not satisfy the molecularclock. A Yule branching process with lognormal priors andempirically defined base frequencies was adopted using thefollowing four calibration points (minimal ages for crowngroups, see also Gottschling et al. 2008): Peridiniales wereestimated at 200 ± 1 Mya (Fensome et al. 1996),Peridiniales excluding Heterocapsaceae were estimated at160 ± 4 Mya based on the first fossil occurrence of(calcareous) coccoid cells with tabulation (Keupp 1984;Keupp and Ilg 1989), Scrippsiella Balech s.l. was estimatedat 70 ± 0.5Mya based on the combination archaeopyle (Strenget al. 2004) as apomorphic trait and the T/Pf clade was esti-mated at 70 ± 0.5 Mya based on the first occurrence ofThoracosphaera Kamptner in the fossil record (Fensome

et al. 1996). For the GTR+Γ substitution model with fourdiscrete categories, we applied an uncorrelated relaxedmolecular clock with a lognormal distribution of rate changes.Starting tree was constructed at random, and the final topologywas estimated by combining five independent chains each of50 million generations, sampling every 10,000th iteration.TRACER v1.6 (http://tree.bio.ed.ac.uk/software/tracer/) wasused to evaluate effective sample size values and to confirmadequate combination of the Markov chain Monte Carlochains with an appropriate burn-in (10%).

Results and discussion

To hinder evolutionary artefacts due to limited and biasedtaxon sample, we first carried out a phylogenetic analysis,including all diatom taxa whose data corresponded with ourcriteria. Accordingly, we constructed a concatenated phyloge-netic analysis with a wide range of endosymbiotic and free-living diatom grades and clades. Our multi-locus and well-resolved phylogenetic tree of diatoms (Fig. S1) confirmedthe existence of three lineages presenting dinophyte endosym-bionts that are only distantly related (Chesnick et al. 1997;Horiguchi and Takano 2006; Pienaar et al. 2007; Takanoet al. 2008; Yamada et al. 2017). We interpret that tertiaryendosymbiosis evolved either early, with many extant free-living diatoms deriving from captured cells, or more likely,resident diatoms were taken up by dinophytes multiple timesindependently (not necessarily in a form of serial replace-ment). Thus, we reject the hypothesis on overall homologyof functional chloroplasts present in Kryptoperidiniaceae. Inthis respect, Kryptoperidiniaceae differ from green algae andland plants, in which plastids have been engulfed only onceand have since been inherited making them homologous.

Two endosymbiotic lineages were recovered asparaphyletic with regard to free-living diatoms, indicatingmultiple independent endosymbiotic events within each sym-biont clade. Therefore, two further phylogenetic analyses werecarried out, including all available sequences (regardless ofthe coverage) within each lineage to elucidate the actualhost-endosymbiont dynamics within each group. As inferredfrom the trees with a focus on distinctive diatom lineages(Figs. 1 and 2), almost every endosymbiont DNA sequencehad its closest relative among free-living algae. Even the res-ident diatoms within particular dinophyte taxa such asDurinskia and Kryptoperidinium did not constitute monophy-le t ic groups . Notably, the ITS sequence of ourKryptoperidinium strain GeoB 459 was almost identical(> 99% similarity) to an ITS sequence (AY574381) derivedfrom a free-living Nitzschia pusillaGrunow (data not shown).This observation rejects any higher-level co-phylogeneticstructure and requires a new interpretation of the origin ofthe dinotoms’ chloroplasts.

Absence of co-phylogeny indicates repeated diatom capture in dinophytes hosting a tertiary endosymbiont 31

Our alternative scenario proposed here describes the re-peated capture of free-living diatoms and their transitorymaintenance, but not necessarily the inheritance by descen-dants over geological times. Evolutionary dynamics inKryptoperidiniaceae thus appears higher than assumed sofar. An almost intact ultrastructure and insignificant genomereduction (Imanian et al. 2010, 2012; Hehenberger et al. 2014)favour our scenario, indicating a rather young and not yetoptimal mutualism. Independent acquisition of diatoms waspostulated for dinophytes already in the previous millennium.This idea of ‘separate endosymbiosis’—that dinophytes havecaptured the same or similar diatom species independently

(Inagaki et al. 2000)—is based on isozyme analysis of twodiatom-harbouring dinophytes. There are extensive enzymaticdifferences between Durinskia baltica (Levander) Carty &El.R.Cox and Kryptoperidinium foliaceum (F.Stein)Er.Lindem., indicating rather divergent adaptations of a hostto its endosymbiont (Whitten and Hoyhome 1986). Similarconclusions supporting the independent acquisition of endo-symbionts have been made based on histone analysis of en-dosymbiotic nuclei (Morris et al. 1993; Chesnick et al. 1997).Such older publications are not cited in the era of molecularphylogenetics anymore (and also not in Yamada et al. 2017),although the studies might have elucidated the accurate

Fig. 1 Polyphyleticendosymbionts of freshwaterUnruhdiniumwith closest relativesamong free-ranging diatoms.Maximum likelihood (ML) tree(−ln = 21,700) of all 37Stephanodiscaceae operationaltaxonomic units available, derivedfrom the comparison ofconcatenated rRNA and plastidsequences. Endosymbiont taxa arehighlighted. Note that we did notobserve any contradictory treetopologies between separatedanalyses of nuclear and plastid loci(not shown). Endosymbionts ofUnruhdinium segregated into twodistinct lineages. Four identicalsequences of different U. penardii(Lemmerm.) Gottschling strainsclustered together (94LBS,.98BPP) and built a well-supported(99LBS, 1.00BPP) sister grouprelationship with free-livingDiscostella nipponica (Skvortzov)A.Tuji & D.M.Williams (strainTNS:AL-5776). A second lineage(100LBS, 1.00BPP) did notinclude endosymbiont sequencesofU. cf. kevei (G.X.Liu&Z.Y.Hu)Gottschling andU. jiulongense (H.Gu) Gottschling only, but alsothose of the free-living strainsL435 and LO4-2. All sequences ofthis lineage were nearly identical.Branch lengths are drawn to scale,with the scale bar indicating thenumber of nt substitutions per site.The numbers on the branches arestatistical support values (MLbootstrap values, values < 50 arenot shown, posterior probabilitiesbelow branches, values < 90 arenot shown; asterisks indicatemaximal support)

32 Žerdoner Čalasan A. et al.

evolutionary scenario. Based on our new phylogenetic in-sights, future research should answer the question,whether a particular dinophyte species always hosts a uniquediatom genotype or if it has the potential to take up differentendosymbionts (as it is indicated for Durinskia andKryptoperidinium in our phylogenetic trees).

Irrespectively of the endosymbionts’ variety, the diatom-harbouring dinophytes are highly selective towards specificgroups and do not recruit diatoms arbitrarily: The

endosymbiont of marine Blixaea, for example, is a part of awell-resolved group within Chaetoceros—probably the larg-est taxon of marine centric diatoms (Rines and Hargraves1988)—and endosymbionts of freshwater Unruhdiniumcluster within freshwater Cyclotella (Alverson et al. 2011), butendosymbionts are neither foundwithin other freshwater speciesof Nitzschia nor of Stephanodiscus Ehrenb. The situation is notas congruent in the diatom clade including endosymbionts ofDurinskia, Galeidinium and Kryptoperidinium, in which

Fig. 2 Polyphyleticendosymbionts of Durinskia,Galeidinium andKryptoperidinium with closestrelatives among free-rangingdiatoms. Maximum likelihood(ML) tree (−ln = 14,945) of all 77Bacillariaceae operationaltaxonomic units available,derived from the comparison ofconcatenated rRNA and plastidsequences. Highlightedendosymbiont taxa were scatteredalong up to ten only distantlyrelated lineages over theBacillariaceae phylogeny. Notethat accessions of neitherDurinskia nor Kryptoperidiniumconstituted monophyletic groups.Within such lineages,endosymbionts were usuallycloser related to free-rangingdiatoms than to otherendosymbionts (though statisticalsupport was not always high).Some endosymbionts werephylogenetically isolated fromfree-living diatoms exhibitinglong branches (e.g. D.kwazulunatalensis N.Yamada,Sym & T.Horig., D. oculata andG. rugatum Tam. & T.Horig.).Branch lengths are drawn to scale,with the scale bar indicating thenumber of nt substitutions persite. The numbers on the branchesare statistical support values (MLbootstrap values, values < 50 arenot shown, posterior probabilitiesbelow branches, values < 90 arenot shown; asterisks indicatemaximal support)

Absence of co-phylogeny indicates repeated diatom capture in dinophytes hosting a tertiary endosymbiont 33

34 Žerdoner Čalasan A. et al.

freshwater and marine taxa are intermingled and cannot be dif-ferentiated. However, they are all assigned to differentNitzschia taxa from marine and freshwater habitatsunderlining the taxonomically selective recognition bythe dinophyte hosts. Cell size may matter in this respect(Litchman et al. 2009), as free-living close relatives ofendosymbionts are rather small diatoms, and they belongto the same taxonomic groups as the endosymbionts fromlarge foraminifers (Lee et al. 1989, 1995). Furthermore, theendosymbiotic cytoplasm of Kryptoperidiniaceae is rich inribosomes (Horiguchi and Pienaar 1994; Tamura et al.2005; Pienaar et al. 2007; Takano et al. 2008), which mayindicate a high metabolic activity. Instead of the well-developed mutualism as present in other photosyntheticallyactive algal groups, the dinophyte host may virtually drain itssymbiont. Overall, the situation rather resembles thedomestication of phycobionts in lichens (Lücking et al. 2009),sacoglossan sea slugs (Händeler et al. 2010) or the facultativelyphototrophic ciliates (Qiu et al. 2016) and foraminifers (Leeet al. 1989, 1995). Similar phenomena are also known fromother dinophytes exhibiting kleptoplastidy (Schnepf andElbrächter 1999; Stoecker 1999; Gast et al. 2007; Takanoet al. 2014).

The question arises whether the last common ancestor ofKryptoperidiniaceae was phototrophic and replaced the pri-mary chloroplast (Keeling 2010) or if it was heterotrophicand developed the ability to photosynthesise secondarily.This question is irrespective of the interpretation that the pri-mary chloroplast of the Peridiniales may have developed to aneyespot in Kryptoperidiniaceae (Dodge 1983; Horiguchi et al.1999; Moestrup and Daugbjerg 2007; Takano et al. 2008).Outgroup comparison does not help answer the question, asphototrophic symbionts (i.e. Zooxanthella K.Brandt) as wellas heterotrophic parasites (i.e. Blastodinium) are found in aclose relationship with Kryptoperidiniaceae (Gottschling and

McLean 2013; Gottschling and Söhner 2013; Kretschmannet al. in press). Nevertheless, it is disputable why an organismwould try to integrate a new endosymbiont, if it has alreadyestablished a permanent functional chloroplast of its own. Theobservations ofKryptoperidinium foliaceumwithout an endo-symbiont (Kempton et al. 2002) also challenge the scenariothat mutualism is entirely obligatory. Thus, the facultativity ofheterotrophy may represent the ancestral condition inKryptoperidiniaceae, as there is no evidence of co-existenceof two types of fully functional chloroplasts either.

Dinophytes hosting a tertiary endosymbiont appear as well-trained taxonomists, which selectively recognise their targetsymbionts (putatively former prey: Zhang et al. 2011b) for theirtight organismal interactions. The evolutionary dynamics ismore complex than previously assumed, and phylogenetic dataindicates that diatom acquisition has taken place repeatedly andmight be ongoing. The complex ecological role of dinophytesemploying a variety of nutrition modes such as phototrophy,heterotrophy and mixotrophy may explain their potential andplasticity in acquiring a great diversity of plastids. The inferredrecent origin is also corroborated by our time estimates (Fig. 3),inwhich theageofmosthost lineages isclearlyolder than thoseofthe corresponding resident diatoms. It is, for example, impossiblefor the last common ancestor of Kryptoperidiniaceae, whichappeared in early Cretaceous, to possess a Nitzschia-like diatomas an endosymbiont, whose stem age has been dated no earlierthan to the late Cretaceous (Sorhannus 2007;Medlin 2015). Theyoung and incomplete realisation of mutualism makesKryptoperidiniaceae an exceptional model for studying the firststeps of chloroplast establishment, as the reduction of the exces-sivemorphological andbiochemical components has not yet tak-enplace.This featuremakes themdifferent fromotherdinophytesalso possessing a tertiary endosymbiont (Gast et al. 2007).

Acknowledgements Financial support was provided by the DeutscheForschungsgemeinschaft (grant GO 1549 10-1) and the MünchenerUniversitätsgesellschaft. We thank Nina Simanovic for improving theEnglish text, and the Scientific Committee of the 11th InternationalConference on Modern and Fossil Dinoflagellates for awarding the firstauthor of this study with the Best Young Scientist Oral Presentation.

References

Altschul, S. F., Madden, T. L., Schäffer, A. A., Zhang, J., Zhang, Z.,Miller, W., & Lipman, D. J. (1997). Gapped BLAST and PSI-BLAST: a new generation of protein database search programs.Nucleic Acids Research, 25, 3389–3402.

Alverson, A. J., Jansen, R. K., & Theriot, E. C. (2007). Bridging theRubicon: phylogenetic analysis reveals repeated colonizations ofmarine and fresh waters by thalassiosiroid diatoms. MolecularPhylogenetics and Evolution, 45, 193–210.

Alverson, A. J., Beszteri, B., Julius, M. L., Theriot, E. C. (2011). Themodel marine diatom Thalassiosira pseudonana likely descendedfrom a freshwater ancestor in the genus Cyclotella. BMCEvolutionary Biology, 11, 125.

�Fig. 3 Cretaceous diversification of dinophytes hosting a diatomendosymb ion t . Da t ed phy logeny o f Pe r i d in i a l e s , w i t hKryptoperidiniaceae fully uncollapsed (ultrametric maximum cladecredibility tree with node ages from the Bayesian uncorrelatedlognormal analysis). Median rate is given in units of substitutions permillion years (including 95% confidence intervals), and branches arecoloured according to the estimated evolutionary rates. Absolute agesare in million years, and epochs are indicated. The numbers on thebranches are statistical support values (Bayesian posterior probabilities,values < .90 are not shown; asterisks indicate maximal support). Note thatKryptoperidiniaceae were monophyletic (.96BPP) and started to diversifyat the Jurassic/Cretaceous boundary. Durinskia (.98BPP) was firstbranching, diversifying since the mid Cretaceous and built a sistergroup relationship to the remaining Kryptoperidiniaceae (.94BPP).Stem node of Unruhdinium (1.00BPP) is subsequently also old, but theextant members did not start to diversify before the late Palaeogene.Among the remaining lineages of Kryptoperidiniaceae, Galeidinium(.99BPP) and Kryptoperidinium appeared closely related (.99BPP), anddiversification of their individual lineages did not take place before theK/Pg boundary

Absence of co-phylogeny indicates repeated diatom capture in dinophytes hosting a tertiary endosymbiont 35

Burki, F. (2014). The eukaryotic tree of life from a global phylogenomicperspective. Cold Spring Harbor Perspectives in Biology, 6,a016147.

Chesnick, J. M., Kooistra, W. H. C. F., Wellbrock, U., & Medlin, L. K.(1997). Ribosomal RNA analysis indicates a benthic pennate diatomancestry for the endosymbionts of the dinoflagellates Peridiniumfoliaceum and Peridinium balticum (Pyrrhophyta). Journal ofEukaryotic Microbiology, 44, 314–320.

Dodge, J. D. (1966). The Dinophyceae. In M. B. E. Godward (Ed.), Thechromosomes of the algae (pp. 96–115). New York: St. Martin’sPress.

Dodge, J. D. (1971). A dinoflagellate with both a mesokaryotic and aeukaryotic nucleus. I. Fine structure of the nuclei. Protoplasma, 73,145–157.

Dodge, J. D. (1983). The functional and phylogenetic significance ofdinoflagellate eyespots. Biosystems, 16, 259–267.

Dodge, J. D. (1989). Phylogenetic relationship of dinoflagellates and theirplastids. In J. C. Green, B. S. C. Leadbeater, & W. L. Diver (Eds.),The chromophyte algae. Problems and perspectives (pp. 205–227).Oxford: Clarendon Press.

Dorrell, R. G., & Howe, C. J. (2015). Integration of plastids with theirhosts: lessons learned from dinoflagellates. Proceedings of theNational Academy of Science of the United States of America, 112,10247–10254.

Drummond, A. J., Suchard, M. A., Xie, D., & Rambaut, A. (2012).Bayesian phylogenetics with BEAUti and the BEAST 1.7.Molecular Biology and Evolution, 29, 1969–1973.

Fensome, R. A., MacRae, R. A., Moldowan, J. M., Taylor, F. J. R., &Williams, G. L. (1996). The early Mesozoic radiation of dinoflagel-lates. Paleobiology, 22, 329–338.

Fensome, R. A., Saldarriaga, J. F., & Taylor, F. J. R. (1999).Dinoflagellate phylogeny revisited: reconciling morphological andmolecular based phylogenies. Grana, 38, 66–80.

Figueroa, R. I., Bravo, I., Fraga, S., Garces, E., & Llaveria, G. (2009). Thelife history and cell cycle of Kryptoperidinium foliaceum, a dinofla-gellate with two eukaryotic nuclei. Protist, 160, 285–300.

Gagat, P., Bodył, A., Mackiewicz, P., & Stiller, J. W. (2014). Tertiaryplastid endosymbioses in dinoflagellates. In Löffelhardt (Ed.),Endosymbiosis (pp. 233–290). Vienna: Springer.

Garcia-Cuetos, L., Moestrup, Ø., Hansen, P. J., & Daugbjerg, N. (2010).The toxic dinoflagellate Dinophysis acuminata harbors permanentchloroplasts of cryptomonad origin, not kleptochloroplasts.HarmfulAlgae, 9, 25–38.

Gast, R. J., Moran, D. M., Dennett, M. R., & Caron, D. A. (2007).Kleptoplasty in an Antarctic dinoflagellate: caught in evolutionarytransition? Environmental Microbiology, 9, 39–45.

Gottschling, M., & McLean, T. I. (2013). New home for tiny symbionts:dinophytes determined as Zooxanthella are Peridiniales and distant-ly related to Symbiodinium.Molecular Phylogenetics and Evolution,67, 217–222.

Gottschling, M., Renner, S. S., Meier, K. J. S., Willems, H., & Keupp, H.(2008). Timing deep divergence events in calcareous dinoflagel-lates. Journal of Phycology, 44, 429–438.

Gottschling,M., & Söhner, S. (2013). An updated list of generic names inthe Thoracosphaeraceae. Microorganisms, 1, 122–136.

Gottschling, M., Söhner, S., Zinßmeister, C., John, U., Plötner, J.,Schweikert, M., Aligizaki, K., & Elbrächter, M. (2012).Delimitation of the Thoracosphaeraceae (Dinophyceae), includingthe calcareous dinoflagellates, based on large amounts of ribosomalRNA sequence data. Protist, 163, 15–24.

Gouy, M., Guindon, S., & Gascuel, O. (2010). SeaView version 4: amultiplatform graphical user interface for sequence alignment andphylogenetic tree building. Molecular Biology and Evolution, 27,221–224.

Gu, H., Kirsch, M., Zinßmeister, C., Söhner, S., Meier, K. J. S., Liu, T., &Gottschling, M. (2013). Waking the dead: morphological and

molecular characterization of extant †Posoniella tricarinelloides(Thoracosphaeraceae, Dinophyceae). Protist, 164, 583–597.

Händeler, K., Wägele, H., Wahrmund, U., Rüdinger, M., & Knoop, V.(2010). Slugs’ last meals: molecular identification of sequesteredchloroplasts from different algal origins in Sacoglossa(Opisthobranchia, Gastropoda). Molecular Ecology Resources, 10,968–978.

Hansen, G., Daugbjerg, N., &Henriksen, P. (2003). Comparative study ofGymnodinium mikimotoi and Gymnodinium aureolum, comb. nov.(= Gyrodinium aureolum) based on morphology, pigment composi-tion, and molecular data. Journal of Phycology, 36, 394–410.

Harper, J. T., Waanders, E., & Keeling, P. J. (2005). On the monophyly ofchromalveolates using a six-protein phylogeny of eukaryotes.International Journal of Systematic and EvolutionaryMicrobiology, 55, 487–496.

Hehenberger, E., Imanian, B., Burki, F., &Keeling, P. J. (2014). Evidencefor the retention of two evolutionary distinct plastids in dinoflagel-lates with diatom endosymbionts. Genome Biology and Evolution,6, 2321–2334.

Horiguchi, T., Kawai, H., Kubota, M., Takahashi, T., & Watanabe, M.(1999). Phototactic responses of four marine dinoflagellates withdifferent types of eyespot and chloroplast. Phycological Research,47, 101–107.

Horiguchi, T., & Pienaar, R. N. (1994). Ultrastructure and ontogeny of anew type of eyespot in dinoflagellates. Protoplasma, 179, 142–150.

Horiguchi, T., & Takano, Y. (2006). Serial replacement of a diatom en-dosymbiont in the marine dinoflagellate Peridinium quinquecorne(Peridiniales, Dinophyceae). Phycological Research, 54, 193–200.

Imanian, B., & Keeling, P. J. (2007). The dinoflagellates Durinskiabaltica andKryptoperidinium foliaceum retain functionally overlap-ping mitochondria from two evolutionarily distinct lineages. BMCEvolutionary Biology, 7, 172.

Imanian, B., Pombert, J. F., Dorrell, R. G., Burki, F., & Keeling, P. J.(2012). Tertiary endosymbiosis in two dinotoms has generated littlechange in the mitochondrial genomes of their dinoflagellate hostsand diatom endosymbionts. PLoS One, 7, e43763.

Imanian, B., Pombert, J. F., & Keeling, P. J. (2010). The complete plastidgenomes of the two ‘dinotoms’ Durinskia baltica andKryptoperidinium foliaceum. PLoS One, 5, e10711.

Inagaki, Y., Dacks, J. B., Doolittle, W. F., Watanabe, K. I., & Ohama, T.(2000). Evolutionary relationship between dinoflagellates bearingobligate diatom endosymbionts: insight into tertiary endosymbiosis.International Journal of Systematic and Evolutionary Microbiology,50, 2075–2081.

Janouškovec, J., Gavelis, G. S., Burki, F., Dinh, D., Bachvaroff, T. S.,Gornik, S. G., et al. (2017). Major transitions in dinoflagellate evo-lution unveiled by phylotranscriptomics. Proceedings of theNational Academy of Science of the United States of America, 114,171–180.

Jeffrey, S. W. (1989). Chlorophyll c pigments and their distribution in thechromophyte algae. In J. C. Green, B. S. C. Leadbeater, & W. L.Diver (Eds.), The chromophyte algae. Problems and perspectives(pp. 13–36). Oxford: Clarendon Press.

Kamikawa, R., Tanifuji, G., Kawachi, M., Miyashita, H., Hashimoto, T.,& Inagaki, Y. (2015). Plastid genome-based phylogeny pinpointedthe origin of the green-colored plastid in the dinoflagellateLepidodinium chlorophorum. Genome Biology and Evolution, 7,1133–1140.

Katoh, K., & Standley, D. M. (2013). MAFFT multiple sequence align-ment software version 7: improvements in performance and usabil-ity. Molecular Biology and Evolution, 30, 772–780.

Keeling, P. J. (2004). Diversity and evolutionary history of plastids andtheir hosts. American Journal of Botany, 91, 1481–1493.

Keeling, P. J. (2010). The endosymbiotic origin, diversification and fateof plastids. Philosophical Transactions of the Royal Society B:Biological Sciences, 365, 729–748.

36 Žerdoner Čalasan A. et al.

Keeling, P. J., Burki, F., Wilcox, H. M., Allam, B., Allen, E. E., Amaral-Zettler, L. A., et al. (2014). The marine microbial eukaryote tran-scriptome sequencing project (MMETSP): illuminating the func-tional diversity of eukaryotic life in the oceans through tran-scriptome sequencing. PLoS Biology, 12, e1001889.

Kempton, J. W., Wolny, J., Tengs, T., Rizzo, P., Morris, R., Tunnell, J.,et al. (2002). Kryptoperidinium foliaceum blooms in SouthCarolina: a multi-analytical approach to identification. HarmfulAlgae, 1, 383–392.

Keupp, H. (1984). Revision der kalkigen Dinoflagellaten-Zysten G.DEFLANDREs 1948. Palaeontologische Zeitschrift, 58, 9–31.

Keupp, H., & Ilg, A. (1989). Die kalkigen Dinoflagellaten im Ober-Callovium und Oxfordium der Normandie/Frankreich. BerlinerGeowissenschaftliche Abhandlungen, A106, 165–206.

Kretschmann, J., Žerdoner Čalasan, A., Gottschling, M. (in press).Molecular phylogenetics of dinophytes harboring diatoms as endo-symbionts (Kryptoperidiniaceae, Peridiniales), with evolutionary in-terpretations and a focus on the identity of Durinskia oculata fromPrague. Molecular Phylogenetics and Evolution.

Leander, B. S., & Keeling, P. J. (2004). Early evolutionary history ofdinoflagellates and apicomplexans (Alveolata) as inferred fromhsp90 and actin phylogenies. Journal of Phycology, 40, 341–350.

Lee, J. J., McEnery, M. E., Ter Kuile, B., Erez, J., Röttger, R., Rockwell,R. F., et al. (1989). Identification and distribution of endosymbioticdiatoms in larger foraminifera. Micropaleontology, 35, 353–366.

Lee, J. J., Morales, J., Symons, A., & Hallock, P. (1995). Diatomsymbionts in larger foraminifera from Caribbean hosts. MarineMicropaleontology, 26, 99–105.

Lee, M. A., Faria, D. G., Han, M. S., Lee, J., & Ki, J. S. (2013).Evaluation of nuclear ribosomal RNA and chloroplast gene markersfor the DNA taxonomy of centric diatoms. Biochemical Systematicsand Ecology, 50, 163–174.

Li, C. L., Ashworth, M. P., Witkowski, A., Dabek, P., Medlin, L. K.,Kooistra, W. H. C. F., et al. (2015). New insights intoPlagiogrammaceae (Bacillariophyta) based on multigene phyloge-nies and morphological characteristics with the description of a newgenus and three new species. PLoS One, 10, e0139300.

Litchman, E., Klausmeier, C. A., & Yoshiyama, K. (2009). Contrastingsize evolution in marine and freshwater diatoms. Proceedings of theNational Academy of Science of the United States of America, 106,2665–2670.

Lücking, R., Lawrey, J. D., Sikaroodi, M., Gillevet, P. M., Chaves, J. L.,Sipman, H. J., & Bungartz, F. (2009). Do lichens domesticatephotobionts like farmers domesticate crops? Evidence from a previ-ously unrecognized lineage of filamentous cyanobacteria. AmericanJournal of Botany, 96, 1409–1418.

McEwan, M. L., & Keeling, P. J. (2004). HSP90, tubulin and actin areretained in the tertiary endosymbiont genome of Kryptoperidiniumfoliaceum. Journal of Eukaryotic Microbiology, 51, 651–659.

Medlin, L. K. (2015). A timescale for diatom evolution based on fourmolecular markers: reassessment of ghost lineages and major stepsdefining diatom evolution. Vie et Milieu, 65, 219–238.

Medlin, L. K., & Fensome, R. A. (2013). Dinoflagellate macroevolution:some considerations based on an integration of molecular, morpho-logical and fossil evidence. In J. M. Lewis, F. Marret, & L. Bradley(Eds.), Biological and geological perspectives of dinoflagellates(pp. 263–274). Geological Society: London.

Moestrup, Ø., & Daugbjerg, N. (2007). On dinoflagellate phylogeny andclassification. In J. Brodie & J. Lewis (Eds.),Unravelling the algae,the past, present, and future of algal systematics (pp. 215–230).CRC Press: Boca Raton.

Morden, C. W., & Sherwood, A. R. (2002). Continued evolutionary sur-prises among dinoflagellates. Proceedings of the National Academyof Science of the United States of America, 99, 11558–11560.

Moreno Díaz de la Espina, S., Alverca, E., Cuadrado, A., & Franca, S.(2005). Organization of the genome and gene expression in a

nuclear environment lacking histones and nucleosomes: the amaz-ing dinoflagellates. European Journal of Cell Biology, 84, 137–149.

Morris, R. L., Fuller, C. B., & Rizzo, P. J. (1993). Nuclear basic proteinsfrom the binucleate dinoflagellate Peridinium foliaceum(Pyrrophyta). Journal of Phycology, 29, 342–347.

Nosenko, T., Lidie, K. L., Van Dolah, F. M., Lindquist, E., Cheng, J. F., &Bhattacharya, D. (2006). Chimeric plastid proteome in the Florida“red tide” dinoflagellate Karenia brevis. Molecular Biology andEvolution, 23, 2026–2038.

Okamoto, N., & Keeling, P. J. (2014). A comparative overview of theflagellar apparatus of dinoflagellate, perkinsids and colpodellids.Microorganisms, 2, 73–91.

Patron, N. J., Waller, R. F., & Keeling, P. J. (2006). A tertiary plastid usesgenes from two endosymbionts. Journal of Molecular Biology, 357,1373–1382.

Pienaar, R. N., Sakai, H., & Horiguchi, T. (2007). Description of a newdinoflagellate with a diatom endosymbiont, Durinskia capensis sp.nov. (Peridiniales, Dinophyceae) from South Africa. Journal ofPlant Research, 120, 247–258.

Pochon, X., Putnam, H. M., Burki, F., & Gates, R. D. (2012). Identifyingand characterizing alternative molecular markers for the symbioticand free-living dinoflagellate genus Symbiodinium. PLoS One, 7,e29816.

Price, D. C., & Bhattacharya, D. (2017). Robust Dinoflagellata phyloge-ny inferred from public transcriptome databases. Journal ofPhycology, 53, 725–729.

Qiu, D., Huang, L., & Lin, S. (2016). Cryptophyte farming by symbioticciliate host detected in situ. Proceedings of the National Academy ofScience of the United States of America, 113, 12208–12213.

Rill, R. L., Livolant, F., Aldrich, H. C., & Davidson, M. W. (1989).Electron microscopy of liquid crystalline DNA: direct evidence forcholesteric-like organization of DNA in dinoflagellate chromo-somes. Chromosoma, 98, 280–286.

Rines, J. E. B., & Hargraves, P. E. (1988). The Chaetoceros Ehrenberg(Bacillariophyceae) flora of Narragansett Bay, Rhode Island, U.S.A.Bibliotheca Phycologica, 79, 1–196.

Rizzo, P. J. (2003). Those amazing dinoflagellate chromosomes. CellResearch, 13, 215–217.

Ronquist, F., Teslenko, M., van der Mark, P., Ayres, D. L., Darling, A.,Hohna, S., et al. (2012). MrBayes 3.2: efficient Bayesian phyloge-netic inference and model choice across a large model space.Systematic Biology, 61, 539–542.

Schnepf, E., & Elbrächter, M. (1988). Cryptophycean-like doublemembranebound chloroplast in the dinoflagellate, DinophysisEhrenb.: Evolutionary, phylogenetic and toxicological implications.Botanica Acta, 101, 196–203.

Schnepf, E., & Elbrächter, M. (1999). Dinophyte chloroplasts and phy-logeny—a review. Grana, 38, 81–97.

Sorhannus, U. (2007). A nuclear-encoded small-subunit ribosomal RNAtimescale for diatom evolution. Marine Micropaleontology, 65, 1–12.

Stamatakis, A. (2014). RAxML version 8: a tool for phylogenetic analy-sis and post-analysis of large phylogenies. Bioinformatics, 30,1312–1313.

Stoecker, D. K. (1999). Mixotrophy among dinoflagellates. Journal ofEukaryotic Microbiology, 46, 397–401.

Streng, M., Hildebrand-Habel, T., & Willems, H. (2004). A proposedclassification of archeopyle types in calcareous dinoflagellate cysts.Journal of Paleontology, 78, 456–483.

Takano, Y., Hansen, G., Fujita, D., &Horiguchi, T. (2008). Serial replace-ment of diatom endosymbionts in two freshwater dinoflagellates,Peridiniopsis spp. (Peridiniales, Dinophyceae). Phycologia, 47,41–53.

Takano, Y., Yamaguchi, H., Inouye, I., Moestrup, Ø., & Horiguchi, T.(2014). Phylogeny of five species of Nusuttodinium gen. nov.

Absence of co-phylogeny indicates repeated diatom capture in dinophytes hosting a tertiary endosymbiont 37

(Dinophyceae), a genus of unarmoured kleptoplastidic dinoflagel-lates. Protist, 165, 759–778.

Tamura, M., Shimada, S., & Horiguchi, T. (2005). Galeidinium rugatumgen. et sp. nov. (Dinophyceae), a new coccoid dinoflagellate with adiatom endosymbiont. Journal of Phycology, 41, 658–671.

Taylor, F. J. R. (1980). On dinoflagellate evolution. Biosystems, 13, 65–108.

Taylor, F. J. R., Hoppenrath, M., & Saldarriaga, J. F. (2008).Dinoflagellate diversity and distribution. Biodiversity andConservation, 17, 407–418.

Theriot, E. C., Ashworth, M. P., Nakov, T., Ruck, E., & Jansen, R. K.(2015). Dissecting signal and noise in diatom chloroplast proteinencoding genes with phylogenetic information profiling.Molecular Phylogenetics and Evolution, 89, 28–36.

Tillmann, U., Gottschling, M., Nézan, E., Krock, B., & Bilien, G. (2014).Morphological and molecular characterization of three newAzadinium species (Amphidomataceae, Dinophyceae) from theIrminger Sea. Protist, 165, 417–444.

Tomas, R., & Cox, E. (1973). Observations on symbiosis of Peridiniumbalticum and its intracellular alga. 1. Ultrastructure. Journal ofPhycology, 9, 304–323.

Whitten, D. J., & Hoyhome, B. A. (1986). Comparative electrophoreticanalysis of two binucleate dinoflagellates. Journal of Phycology, 22,348–352.

Wisecaver, J. H., & Hackett, J. D. (2011). Dinoflagellate genome evolu-tion. Annual Review of Microbiology, 65, 369–387.

Yamada, N., Sym, S. D., & Horiguchi, T. (2017). Identification of highly-divergent diatom derived chloroplasts in dinoflagellates, including adescription of Durinskia kwazulunatalensis sp. nov. (Peridiniales,Dinophyceae). Molecular Biology and Evolution, 34, 1335–1351.

You, X. J., Luo, Z. H., Su, Y. P., Gu, L., & Gu, H. (2015). Peridiniopsisjiulongensis, a new freshwater dinoflagellate with a diatom endo-symbiont from China. Nova Hedwigia, 101, 313–326.

Zhang, Q., Liu, G.-X., & Hu, Z.-Y. (2011a). Durinskia baltica(Dinophyceae), a newly recorded species and genus from China,and its systematics. Journal of Systematics and Evolution, 49,476–485.

Zhang, Q., Liu, G.-X., & Hu, Z.-Y. (2011b). Morphological differencesand molecular phylogeny of freshwater blooming species,Peridiniopsis spp. (Dinophyceae) from China. European Journalof Protistology, 47, 149–160.

38 Žerdoner Čalasan A. et al.