Accepted Manuscript Title: Scale evolution, sequence phylogeny, and taxonomy of thaumatomonad Cercozoa: 11 new species and new genera Scutellomonas, Cowlomonas, Thaumatospina and Ovaloplaca Author: Josephine M. Scoble Thomas Cavalier-Smith PII: DOI: Reference: S0932-4739(13)00075-8 http://dx.doi.org/doi:10.1016/j.ejop.2013.12.005 EJOP 25316 To appear in: Received date: Revised date: Accepted date: 6-10-2013 16-12-2013 17-12-2013 Please cite this article as: Scoble, J.M., Cavalier-Smith, T.,Scale evolution, sequence phylogeny, and taxonomy of thaumatomonad Cercozoa: 11 new species and new genera Scutellomonas, Cowlomonas, Thaumatospina and Ovaloplaca, European Journal of Protistology (2013), http://dx.doi.org/10.1016/j.ejop.2013.12.005 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Scale evolution, sequence phylogeny, and taxonomy of thaumatomonad Cercozoa: 11 new species and new genera Scutellomonas, Cowlomonas, a us cr ip Josephine M. Scoblea,* , Thomas Cavalier‐Smitha t Thaumatospina and Ovaloplaca pt ed M an Department of Zoology, University of Oxford, South Parks Road, Oxford. OX1 3PS, UK ce _______________ *Corresponding author. Tel.: +44 1865 281906; fax: +44 1865 281310. Ac E‐mail address: josephine.scoble@zoo.ox.ac.uk (J. M. Scoble) Page 1 of 92 2 Abstract We describe 11 new species of Thaumatomonadida using light and electron microscopy and rDNA gene sequences (18S, ITS1, 5.8S, ITS2). We found clear us cr ip t distinctions between major clades in molecular and morphological traits that support now splitting Thaumatomastix into three genera: new marine genera Ovaloplaca (oval plate‐scales) and Thaumatospina (triangular plate‐scales), both with distinctive radially‐symmetric bobbin‐based spine‐scales, restricting Thaumatomastix to freshwater species with putatively non‐homologous eccentric‐spine scales and thicker triangular plate‐scales. New genus Scutellomonas lacks spine‐scales, having oval plate‐ an scales with deeply‐dished upper tier as in Ovaloplaca, with which it forms a clade having short/absent anterior cilium. Cowlomonas gen. n. is possibly naked. We describe two new Allas species, two new Thaumatomonas, and one new Reckertia M species, and transfer R. hindoni to Thaumatomonas. Triangular‐scaled Reckertia has varied plate‐scales and ciliary scales. Thaumatomonas rDNA trees reveal two clades: ed zhukovi/seravini (predominantly triangular scales); coloniensis/oxoniensis/lauterborni/constricta/solis (scales mostly oval). We hypothesize that the ancestor of Thaumatomonadidae had radially‐symmetric bobbin‐ pt based spine‐scales and triangular plate‐scales, bobbin‐based spine‐scales being lost in one lineage and eccentric‐spine scales evolving in Thaumatomastix. Bobbin‐based ce spine‐scales arguably evolved from triangular plate‐scales and single‐tier ciliary scales (Ovaloplaca and Reckertia only) from plate‐scale rudiments. We present a unified Ac scheme for scale evolution and development in Imbricatea. Keywords: Imbricatea; Silica scale evolution; Spongomonadida; Thaumatomonadida;, Thaumatomastix; Zoelucasida Page 2 of 92 3 Introduction For over 700 million years mineralised cell surface scales have featured in protist evolution, the oldest known being phosphatic (Cohen and Knoll 2011). Thaumatomonads belonging to phylum Cercozoa are colourless gliding biciliates us cr ip t characterised by two‐tier silicified scales made in silica‐deposition vesicles partially embedded in giant mitochondria (Ota et al. 2012). Thaumatomonadida often feed by extending filose pseudopodia from a ventral groove and are grouped with the filose, purely amoeboid Euglyphida as class Imbricatea as both have often imbricate silica scales (Cavalier‐Smith and Chao 2003), although the single‐tier euglyphid scales are made in the Golgi region. Ribosomal DNA sequence evidence broadened Imbricatea to an include also several scaleless taxa, e.g. Spongomonadida and Marimonadida (Howe et al. 2011a). M Recently, Thaumatomonadida were grouped with Rotosphaerida, filose silica‐ scaled non‐flagellates (families Pinaciophoridae, Clathrellidae, Rabdiophryidae, Rabdiocystidae, Luffisphaeridae), in a new superorder Placofila characterised by two‐ ed tier plate scales (Cavalier‐Smith and Chao 2012), but no rotosphaerid sequence data are available to verify that postulated relationship. The biciliate Spongomonadida, with pt some ultrastructural similarities to Thaumatomonadida (Cavalier‐Smith and Karpov 2012) and their likely closest relatives amongst sequenced taxa (Howe et al. 2011a), ce are grouped with Perlofilida (unsequenced filose non‐flagellates Pompholyxophrys, Acanthoperla) as superorder Perlatia, which in contrast to Thaumatomonadida and Ac Euglyphida secrete silicified perles instead of scales (Cavalier‐Smith and Chao 2012). Placofila and Perlatia jointly constitute the imbricate subclass Placoperla (Cavalier‐ Smith and Chao 2012); except for the secondarily scale‐less thaumatomonad, Esquamula lacrimiformis (related to the scale‐bearing Peregrinia clavideferens: Shiratori et al. 2012), all currently known Placoperla secrete either two‐tiered silica scales or globular silica perles. The other imbricate subclass Placonuda also has two superorders: Euglyphia with conspicuous imbricated single‐tier plate scales and Nudisarca (Marimonadida, Nudifilidae: the only known imbricates other than Esquamula without silicified surface Page 3 of 92 4 structures) (Cavalier‐Smith and Chao 2012). At present Euglyphia include only the non‐ flagellate filose testate Euglyphida, but a non‐filose biciliate, Zoelucasa, was recently discovered with scales so closely similar to those of the euglyphid Sphenodenia (Nicholls 2012a) that it might possibly belong in Euglyphia as sister to euglyphids. t Zoelucasa, with ciliary pit and cilia arranged as in thaumatomonads, swims rather than us cr ip glides and is the only scaly imbricate flagellate known outside Thaumatomonadida, and is here placed in a new imbricate order, Zoelucasida, seemingly phenotypically intermediate between thaumatomonads and euglyphids. The closest relatives to the putatively ancestrally scaly Imbricatea are Thecofilosea, which never have scales and are grouped with Imbricatea as superclass Ventrifilosa (Cavalier‐Smith and Karpov an 2012). Most thaumatomonad species have been described solely from whole‐mount electron micrographs of scale morphology in environmental samples. Thus, for most, M the appearances of living cells or sequences and their trophic requirements are completely unknown. Recently, clonal cultures, sequencing, and electron microscopy ed were used to establish new species of Thaumatomonas, a new genus Peregrinia, and to segregate Reckertia from Thaumatomastix (Howe et al. 2011a). That revision restricted Thaumatomastix to species having both spine‐scales and plate‐scales. Even pt then it was clear that the revised Thaumatomastix still included marine species with very different plate‐scales: either triangular or oval. Recently Nicholls (2012b) found ce that the spine‐scales of two freshwater Thaumatomastix species are so different from those of marine species that we here consider them separate genera with non‐ Ac homologous spine‐scales. To clarify further thaumatomonad taxonomy, phylogeny, and scale evolution we studied 34 new clonal cultures by electron and light microscopy and sequencing. We show that marine Thaumatomastix comprise two morphologically and genetically distinct clades with triangular versus oval plate‐scales, here made new genera distinct from freshwater Thaumatomastix, which include the not yet sequenced type species Thaumatomastix setifera (Lauterborn 1899). We establish new genus Scutellomonas for a new clade of thaumatomonads with oval plate‐scales but no spine‐scales. Page 4 of 92 5 We describe new species of Allas, Thaumatomonas, and Reckertia, and transfer Reckertia hindoni Nicholls (2012b) to Thaumatomonas. We resequenced three Thaumatomonas cultures whose labelling had confused thaumatomonad taxonomy (Howe et al. 2011a) and definitively dispel those confusions. Because Thaumatomonas t and Allas 18S rDNA sequences are so similar, we sequenced ITS regions of 31 us cr ip ribosomal (r) DNAs for more phylogenetic resolution and clearer species demarcation, using ITS2 rDNA secondary structure and ITS/5.8S rDNA region trees, which confirms the previously questionable phylogenetic distinctness of Thaumatomonas and Allas. an Material and Methods Culture isolation M We used clean plastic sheets to scrape soil or plastic test tubes to collect liquid (1 ‐ 5 g) samples from soil, marine littoral sand, seawater, or freshwater, which were placed in sterile 90 mm Petri dishes and drowned in media with a sterile (boiled) ed barley grain. Volvic® (Danone) was used for freshwater and soil samples. Artificial Seawater for Protists (ASWP; Culture Collection of Algae and Protozoa (CCAP) recipe: pt www.ccap.ac.uk) was used for marine samples. Uniprotist clonal cultures were obtained from these crude cultures by single‐cell picking and/or serial dilution. Strains ce BKARH9, GSPB9, HANTSF8 and MLTB12, were provided by K. Vickerman. Surviving strains representing new species have been placed in CCAP (accession numbers in Ac diagnoses). Strain YPF609 (Nies Culture collection Japan: NIES‐2378) was studied by A. Yabuki who kindly gave us his sequence and micrographs. Microscopy Phase contrast and differential interference contrast (DIC) light microscopy. Cultures were viewed and manipulated using phase contrast X20 and X40 objectives (Olympus IX70: Olympus CK40). Live high definition video recordings were made using a Sony HDV 1080i Handycam® mounted on a Nikon Eclipse 80i microscope with a water immersion x60 DIC objective lens (NA 1.0). Videos were uploaded to computer and cell measurements taken from still frames. Stills were extracted using Final Cut Express HD Page 5 of 92 6 3.5.1 and Adobe Photoshop CS4 used to make the figures. Sometimes it was difficult to measure dimensions if cells did not lie completely flat (tilting causes foreshortening), so cells were measured at several stages of movement for accuracy, and results for each cell averaged. Electron microscopy. Whole‐mount formvar‐coated 200 mesh t copper metal grids were viewed in an FEI Technai 12 electron microscope. Cells were us cr ip fixed in glutaraldehyde diluted to 2.5% in freshwater or salt water media, dried on the grid, then washed with deionised water. YPF609 was fixed and photographed via SEM by A. Yabuki (Yabuki and Ishida 2011). DNA extraction and sequencing DNA was extracted as soon as possible after initial isolation, when cultures an were in their first ‘bloom’ at the height of the growth curve. Cells were filtered through sterile WhatmanTM GF/F glass fibre filters and DNA extracted using UltraClean® soil M isolation kits (Mo Bio Laboratories, Camb Bio, UK). Eukaryote‐wide* and cercozoan‐ specific** PCR primers (Bass and Cavalier‐Smith 2004; Karpov et al. 2006) (from InvitrogenTM UK) were used for amplifying rDNA: for 18S we used either 25F* ‐ 1801R* ed or 25F* ‐ 1256R**; for ITS1+ITS2 we used 1259F** ‐ 28SR1*, in 25 µl volumes. PCR programs were: 95oC for 5 min (initial denaturation) followed by 35 cycles: 95oC for 32 pt s, annealing 70oC (**) or 68oC (*) (depending on primers, shown by asterisks), extension 72oC for 2.5 min; final extension at 72oC was for 7 min. PCR products were ce electrophoresed on 1% agarose gels and visualised in UV by ethidium bromide. Polyethylene glycol was used to clean single bands (Howe et al. 2011a), but if multiple bands were present the correct size band was cut out and cleaned for sequencing Ac using GFXTM PCR DNA Band Purification Kit (QIAGEN®, GE healthcare UK). Internal sequencing primers were: 18S: 3NDF (5’ GGC AAG TCT GGT GCC AG 3’), and Pre3NDF (5’ CAG CAG GCG CGC AAA TTA CC 3’) (Bass and Cavalier‐Smith 2004). ITS1 and ITS2: PreBF (5’ GTA GGT GAA CCT GCA GAA GGA TC 3’), and 384F (5’ YTB GAT GGT AGT GTA TTG GA 3’) (Dopheide et al. 2009). Sequencing was by Applied Biosystems BigDye® Terminator v. 3.1 and v. 1.1 Cycle Sequencing Kits and Applied Biosystems 3730/3730xl DNA Analyzer (Life Technologies) with Capillary‐system. Page 6 of 92 7 Sequence editing, alignment, and phylogenetic analyses Chromatograms were viewed and edited using Sequence Scanner v. 1.0 of Applied Biosystems. BioEdit was used for contig assembly. Additional 18S rDNA cercozoan sequences were downloaded from NCBI (http://www.ncbi.nlm.nih.gov/). To us cr ip t find closely related 18S and ITS rDNA environmental sequences in NCBI, we used our new sequences as well as published named thaumatomonad sequences as BLAST queries. Sequences were initially aligned using default settings of MAFFT v.6 Online version using secondary structure ‘extremely slow Q‐INS‐I’ setting (Katoh et al. 2009). MacGDE v. 2.4 (http://macgde.bio.cmich.edu/) was used to improve alignments manually. an We made two rDNA alignments. First, a large 18S rDNA alignment containing 273 taxa representing the three classes of infraphylum Monadofilosa closest to M thaumatomonads: Imbricatea, Thecofilosea, and Sarcomonadea. Because the demarcation between Imbricatea and Thecofilosea is poorly resolved, we could not simply assume Thecofilosea to be the correct outgroup for Imbricatea, so it was ed essential to include a broadly representative selection of all major ventrifilosan lineages (Imbricatea, Thecofilosea) and use sarcomonads as an unambiguous outgroup pt for them. This alignment contained many environmental sequences, some mislabeled ‘thaumatomonad’ and others genuinely thaumatomonad. As preliminary analysis ce suggested that some environmental sequences labeled thaumatomonad were actually sarcomonads, we included 73 sarcomonad sequences to establish their position. Ac Unlike Howe et al. (2011a) we excluded the most distant monadofilosan class (Metromonadea), restricting our analysis to Ventrifilosa plus their immediate outgroup Sarcomonadea. By excluding more distantly related Cercozoa, we could align virtually the whole of 18S rDNA with confidence. Our second alignment was of the ITS1, 5.8S and ITS2 rDNA regions representing five thaumatomonad genera. This alignment embraced the last part of the 18S rDNA sequence (~400 nucleotides) and the beginning of 28S (~300 nucleotides) and yielded a more robust branching order for the Thaumatomonas/Allas part of the tree, which is poorly resolved by 18S rDNA alone. This ITS rDNA alignment also includes mislabeled ‘fungal’ environmental clones Page 7 of 92 (Bowden et al. 2004; Hawkes et al. 2011; Waldrop et al. 2006) recognised here as 8 thaumatomonad. Maximum likelihood analyses used the GTRMIX model of RaXML‐7.0.4 (Stamatakis 2006) with 1000 rapid bootstrap resamplings. Two different Bayesian us cr ip t methods were used. MrBayes v. 3.1.2 consensus trees were generated from two chains run for five million generations under a GTR adgamma evolutionary model with four gamma categories and covarion setting. One out of every 1000 trees were kept for analysis, and the first 450 trees excluded as burnin for both 18S and ITS. PhyloBayes v. 3.2 trees used the heterogeneous GTR‐CAT model with default gamma settings (Lartillot et al. 2004) and two chains. Though both chains plateaued, even an after >75,140 cycles maxdiff was 0.50609, so the statistical criterion for convergence in a ‘good’ run was not met; therefore chains were summed separately (as well as together, excluding the first 1288 trees as burnin). As the topology was identical within M the Thaumatomonadida/Spongomonadida clade and within all other major clades for each chain separately and the joint tree and support values closely similar, we ed arbitrarily used chain 1 only based on 73,850 trees for Fig. 1. The key difference between the two chains that accounts for their non‐convergence was in the position of Discomonadida: weakly sister to Imbricatea in chain 1 (Fig. 1) and weakly sister to the pt imbricate subclass Placoperla only (i.e. within imbricates) in chain 2. As the PhyloBayes CAT‐GTR heterogeneous model is evolutionarily more realistic than the homogeneous ce ML and MrBayes models (Lartillot et al. 2004), Figure 1 shows the PhyloBayes tree with Ac support values for all three models mapped onto it. ITS2 fold analysis ITS2 rDNA region was located using an ITS2 annotation program (Koetschan et al. 2009) http://its2.bioapps.biozentrum.uni‐wuerzburg.de/. The clipped ‘annotated’ sequences were folded using the online MFold program (Zuker 2003) http://mfold.rna.albany.edu/?q=mfold/RNA‐Folding‐Form set at 37oC. Each fold was checked by eye within the helix III region for compensatory base changes (CBCs) and hemi‐CBCs (Coleman 2009) compared with closely related strains. Page 8 of 92 GenBank accession numbers: 18S rDNA. KC243105 ‐ KC243105; ITS1, 5.8S rDNA and 9 ITS2 submitted as contiguous sequences, KF577807‐KF577836. us cr ip t Results We studied 23 newly isolated cultures and nine uncharacterised strains from K. Vickerman by sequencing 18S rDNA and transmission electron microscopy (TEM), though sequencing was unsuccessful for three and TEM for several. To clarify decisively Thaumatomonas species‐name confusions discovered by Howe et al. (2011a) we restudied similarly three older cultures (Thaumatomonas sp. from an Vickerman and type cultures of Ts. zhukovi and seravini from Mylnikov). We also sequenced 18S rDNA from two dead unidentified clonal cultures from D. Bass. Fourteen sequences were novel. Thirty‐one cultures were also sequenced across the M internal transcribed spacer regions (ITS1, 5.8S, ITS2) of rDNA. Though our primary goal was to clarify the diversity and phylogeny of named thaumatomonads, we wished also ed to test whether environmental sequences annotated as ‘thaumatomonads’ really are from this group and to clarify the unresolved basal branching of Imbricatea, the class containing thaumatomonads. Because of its large size our monadofilosan 18S rDNA pt tree is split for clarity into Ventrifilosa (Figure 1.1) and the sarcomonad outgroup ce (Figure 1.2), which were consistently separated by all three methods. Support is weak for the tree’s backbone (as previously: Howe et al. 2011a) except for the robust separation of cercomonads from other groups (1, 99%, 1) and the monophyly of Ac Ventrifilosa by ML and MB. Overall phylogeny within Ventrifilosa Within Imbricatea, Placonuda and Placoperla are each reproducibly holophyletic with all three methods (Fig. 1.1). Within Placoperla, Spongomonadida are sister to all Thaumatomonadida in ML and MB (BS 58%, PP 0.86), but in PB spongomonads branch within Thaumatomonadida as sister to Esquamulidae and Peregriniidae (PP 0.61 Fig. 1.1). In ML and MB Thaumatomonadida were consistently holophyletic with Peregriniidae and Thaumatomonadidae sisters, albeit with weak Page 9 of 92 10 support (BS 53%, PP 0.77). Within Placonuda superorder Nudisarca is a clade with PB, but paraphyletic with ML and MB. Imbricatea are holophyletic using PB and ML, but with MB Placonuda branched more deeply (PP 0.56) as sister to all other Ventrifilosa. The MB and ML analyses never grouped Discomonas within Imbricatea, as did MB in t Howe et al. (2011a), or even as sister to imbricates. However, PB showed the us cr ip Discomonas clade either as sister to all other Imbricatea (Fig. 1 chain 1) or (chain 2) as sister to Placoperla only as in the MrBayes tree of Howe et al. (2011a). Only with MB is Ventricleftida sister to Thecofilosea (PP 0.77) in accordance with their classification and the MB tree in Howe et al. (2011a); with PB and ML Ventricleftida are extremely weakly sister to Imbricatea (PP 0.34, BS 5%). an Mataza hastifera (Yabuki and Ishida 2011) is always within Thecofilosea and in both Bayesian trees is sister to Ebriida plus environmental clades eEbriida and AB27503, with moderate support; in ML Matazida are weakly sisters to Cryomonadida M (BS 18%). These trees consistently disagree strongly with both previous contradictory placements of Mataza: within Cryomonadida or as sister to all other Thecofilosea ed (Yabuki and Ichida 2011). In separate 18S rDNA analyses (ML only, not shown as their presence or pt absence did not alter branching within the Thaumatomonadida/Spongomonadida clade) we examined the position of the long‐branch Phaeodaria (Howe et al. 2011a) ce and of the even longer‐branch Sainouroidea (Cavalier‐Smith et al. 2009). When Phaeodaria sequences were added, they were sister to Pseudodifflugia cf. gracilis; Ac Pseudodifflugia cf. gracilis was within Thecofilosea in every tree. When both Phaeodaria and Sainouroidea were added they were mutually sisters, this clade being sister to Pseudodifflugia. However, when Phaeodaria were removed, Sainouroidea moved from Thecofilosea to Imbricatea, weakly sister to Euglyphida. Thaumatomonadida 18S rDNA phylogeny Eight known genera of Thaumatomonadida are in the tree (Fig. 1.1); at least four more are implied by the environmental sequences. The deepest branching of Thaumatomonadida is between Peregriniidae/Esquamulidae (a robust clade) and Thaumatomonadidae (always a clade but with weaker support). Peregrinia Page 10 of 92 11 (freshwater) is sister to the soil sequence, Amb_18_1066 (EF023480) (Lesaulnier et al. 2008), this non‐marine clade being robustly sister to marine Esquamula lacrimiformis (PP1, BS 100%, PP 1). Thaumatomonadidae consist of a single maximally supported freshwater clade (FW in Fig. 1.1) nested within four deeper‐branching marine lineages, t which are distinct on both ML and Bayesian trees. The large freshwater clade us cr ip comprises Thaumatomonas, Allas, soil sequence Amb_18S_1164 (EF023728 and others not shown, EF023966, EF025032, EF023867: Lesaulnier et al. 2008), and another from a paddy field (AB534345: Takada and Morimoto 2010) and is sister to the marine Reckertia‐containing clade. These sister clades both lack spine scales (in strains studied by both sequencing and EM). Three of the four deeper branching lineages of Thaumatomonadidae include species with spine scales that would previously formerly an all have been lumped as ‘Thaumatomastix’ (Howe et al. 2011a), but the fourth (uncultured sequence A15) is of unknown morphology. Thaumatomonas appears M paraphyletic, with Allas sister to the Ts seravini/zhukovi subclade (abbreviations: Thaumatomonas – Ts; Thaumatomastix – Tx; Thaumatospina – Ta). No environmental ed sequences were identified as Thaumatomonas or Allas. Spine‐scaled ‘Thaumatomastix’ are paraphyletic, having two strongly supported clades: Thaumatospina gen. n. with triangular plate‐scales (PP 0.99, BS 90%, pt PP 1), and Ovaloplaca gen. n. with oval plate‐scales and spine‐scales (PP 1, BS 100%, PP 1). Ovaloplaca is sister to a newly described genus, Scutellomonas gen. n. with oval ce plate‐scales only (PP 1, BS 100%, PP 1) and an environmental clone 9‐2.6 (AY620310) (Bass and Cavalier‐Smith 2004) likely to have oval scales. Thus, all oval‐scale Ac thaumatomonads are molecularly more related than any are to the deepest triangular‐ scale Thaumatospina clade. Marine environmental sequence 'A15' (AY620317: Bass et al. 2005) is sister to all Thaumatomonadidae other that the deepest branching Thaumatospina. The two Reckertia species are weakly sister and group robustly with five marine sequences, possibly but not necessarily all Reckertia; three of these unidentified sequences form a robust clade R1 (Fig. 1.1) of sea ice sequences, FN690394, FN690393, FN690392 (Majaneva et al. 2011); EF100294 is from oxygen‐ depleted intertidal marine sediment (Stoeck et al. 2007), and GU385595 from seaweed/bait (Haska et al. 2012). This Reckertia‐containing clade is reproducibly sister Page 11 of 92 12 to the freshwater clade, not to either marine ‘Thaumatomastix’ clade, reinforcing their generic distinction. Thus there are three major distinct thaumatomonad marine clades including known species, all lumped with the freshwater Thaumatomastix prior to Howe et al. (2011a). us cr ip t Identifying environmental DNA clades Two environmental sequences were misannotated as thaumatomonads, from the same soil study (Lesaulnier et al. 2008) from which many previous papers found numerous misannotations affecting other groups (e.g. Cavalier‐Smith and Chao 2012; Cavalier‐Smith and Scoble 2013; Howe et al. 2011a): Amb_18S_1199 (EF023758) is maximally supported as sister to Nudifila, placed below in the new family Nudifilidae: an environmental sequence Elev_18S_823 (EF024516) is not even an imbricate or ventrifilosan sequence, but belongs to the deep‐branching probable glissomonad clade M Te that also includes a sequence (EF023937) even more bizarrely annotated as an apicomplexan eimeriid (Lesaulnier et al. 2008). Figure 1.1 does however reveal three distinct environmental clades lacking cultured representatives that are genuinely sand (AY620317). ed within Thaumatomonadidae, two from soil (AB534345, EF023728), one from marine pt The three core NC1 sequences (Bass and Cavalier‐Smith 2004) are robustly sisters to Marimonadida as Howe et al. (2011a) found, and probably belong in that ce order. However, Figure 1.1 sequence 7‐6.6 (AY620336), which originally grouped weakly with NC1 (Bass and Cavalier‐Smith 2004) or Euglyphida (Bass et al. 2005) or Ac NC3 (Howe et al. 2011a), now consistently groups as sister to a robust clade comprising Nudifilidae, Clautriaviidae and NC2 (robustly sister to Clautriaviidae); below we make this entire clade a new placonudan order Variglissida. Figure 1.1 shows NC3 within Nudisarca (Placonuda) as sister to Marimonadida, NC1 and sequence 9‐2.2. However, in both ML and MB, NC3 is sister to euglyphids and Nudisarca are paraphyletic. Our trees confirm that two core NC4 sequences (AY620348 7‐6.2; AY620314 9‐ 1.4: Bass and Cavalier‐Smith 2004) belong in Thecofilosea, whereas the marine sediment sequence 9‐2.2 (AY620308), originally weakly grouping with NC4 (Bass and Page 12 of 92 13 Cavalier‐Smith 2004) belongs elsewhere. Sequence 9‐2.2 is weakly (PP 0.22) sister to Marimonadida within Nudisarca in the heterogenous PB tree, but weakly sister to a possibly artefactual clade comprising Euglyphida and NC3 (NC3 is within Marimonadida by PB) on the two homogeneous model trees (ML 24%, MB 0.88). No t tree placed 9‐2.2 within glissomonads as it was with no support in Howe et al. (2011a); us cr ip though its precise position is weakly supported, all three methods strongly exclude it from glissomonads. Though our sampling of monadofilosan environmental sequences is more broadly representative than other studies, it is not exhaustive. For example, Fig. 1.1 includes eight environmental sequences related to Cryothecomonas aestivalis, mainly an from the Arctic Ocean and sea ice, except two sister sequences (AB275093, AB275046) obtained from the Japanese east coast from a ‘methane cold seep sediment’ (Takishita et al. 2007); yet these all differ from the 13 Cryothecomonas Arctic environmental DNA M sequences in the tree of Thaler and Lovejoy (2012), emphasising their high diversity in the Arctic. Within Ventricleftida we found a novel clade of marine environmental ed sequences (eVentri) strongly sister to Ventrifissura in PB (but the deepest branch of this clade is one node deeper in ML and MB). pt Two sarcomonad environmental sequences (AB534506 Takada and Morimoto 2010, and EF024287 Lesaulnier et al. 2008) form a consistent clade that is sister to ce both Pansomonadida and Glissomonadida in PB and ML; oddly, with MB two NC7 environmental sequences, (AY620287 and AY620288: Bass et al. 2005) group as their Ac sister instead of within Pansomonadida as with PB, ML and previously (Bass et al. 2005). We name this clade eSarcomonad (Fig 1.1); like clade Te (Howe et al. 2009) it is likely to be morphologically more glissomonad‐like than cercomonad‐like. ITS1, 5.8S, and ITS2 rDNA phylogeny To resolve the relationship between Allas and Thaumatomonas and other thaumatomonads, which was contradictory on PB, ML and MB 18S rRNA trees, we analysed a 1561‐nucleotide alignment of 35 Thaumatomonadidae alone for their ITS1, 5.8S, and ITS2 rDNA sequences, which include ~700 nucleotides from adjacent 18S and 28S rRNA (Fig. 2). Five genera were represented, the tree being rooted between Page 13 of 92 14 Thaumatospina and all other Thaumatomonadidae in accordance with the 18S rDNA tree (Fig. 1.1). In contrast to the 18S rRNA trees Allas and Thaumatomonas were both holophyletic mutually sisters on the MrBayes tree (Fig. 2), which supports their retention as separate genera (Howe et al. 2011a). The positions of named t Thaumatomonas sequences within the two Thaumatomonas clades were exactly the us cr ip same in ML and MB, though three of four misannotated ‘fungal’ clones behaved differently: in MB these three (DQ420715, DQ420716 and JQ666760: Waldrop et al. 2006) were sister to Allas plus Thaumatomonas, whereas in ML this long branch joined the positionally stable short‐branch sequence HM240097 (Hawkes et al. 2011) within Thaumatomonas, as sister to the reproducible Ts coloniensis/oxoniensis/lauterborni/constricta/solis clade; (abbreviated below as C/L/S an referring to the first species of its main sub‐branches). Possibly the MB tree placed the long branch too deeply and all four ‘fungal’ clones are undescribed Thaumatomonas M related to the C/L/S clade; they must be undescribed (three species only as two are almost identical) since Figures 1.1 and 2 together now include all known Thaumatomonas species, assuming that Mylnikov’s T‐1 culture (which grouped closely ed and robustly with Ts constricta on Figure 1.1, but ITS rDNA is not sequenced) was correctly identified as Ts lauterborni. Allas was strongly holophyletic in both 18S pt (1/100%/1) and ITS (0.95/66%) rDNA trees. Relationships amongst Ts ‘vancouveri', Ts coloniensis and Ts oxoniensis are clearer than with 18S rDNA alone. Scutellomonas was ce consistently sister to Reckertia, which contradicts the 18S rRNA tree where it branched one node more deeply; however, the ITS tree is less taxon‐rich near the base, so in this Ac respect might be misleading as it does not include Ovaloplaca, the much closer true sister of Scutellomonas. Phylogeny of the sarcomonad outgroup Glissomonadida and Pansomonadida are consistently sister in MB and ML as in most previous studies. However with PB Pansomonadida branch well within Glissomonadida, albeit with weak support (Fig. 1.2), suggesting that their previous frequent (not invariable) exclusion by some other methods may have been a long‐ branch artefact, thus questioning their status as a separate order. Novel clade 7 (Bass and Cavalier‐Smith 2004) forms three distinct clades, two within pansomonads and Page 14 of 92 one (AY620284) their sister. The pansomonad/AY620284 clade is sister to a robust 15 clade X/Y, previously assumed to be glissomonads (Howe et al. 2011b; Hess and Melkonian 2013), but since our analysis subclade X was identified as a more amoeboid new family Viridiraptoridae (Hess and Melkonian 2013) but probably belongs to the t previously phylogenetically unplaced order Pseudosporida (Cavalier‐Smith 1993); see us cr ip Discussion. With all methods the basal branching order within Glissomonadida differs in a few respects from Howe et al. (2011a,b), possibly owing to our improved alignment including more nucleotides and more extensive outgroups. Even for the most distant outgroup Cercomonadida our trees are better resolved and more consistent than earlier. Cercomonas is a distinct clade in ML, but an paraphyletic with PB and MB. Within Cercomonadidae, Neocercomonas plus Filomonas form a robustly supported clade in all three analyses (PP 0.99, BS 82%, PP 1.0), invalidating the assertion (Brabender et al. 2012) that bootstrap values are too low to (Cavalier‐Smith and Karpov 2012). M support the monophyly of Neocercomonas and its distinction from Cercomonas ed ITS2 rDNA secondary structures of Thaumatomonadida Thaumatomonas exhibited multiple isolates with the same 18S rDNA sequence, pt some with different or additional scale morphologies. Though the ITS rDNA region tree gave better resolution (Fig. 2), especially within the Ts zhukovi/seravini (Z/S) clade, and ce C/L/S clade, their internal branches are still so close together and in some cases poorly resolved, and scale morphologies so variable and not mapping simply onto the tree, Ac that it was hard to use either trees or morphology to make clear boundaries between species. We therefore also used ITS2 rDNA secondary structure to help such demarcation, as Coleman (2009) suggested that organisms differing by at least one compensatory base change (CBC) in helix III would be unable to cross, and therefore are probably separate biological species. Figure 3 shows that ITS2 rDNA secondary structure is generally well conserved across Thaumatomonadidae and has the canonical structure of helices I‐III. Structure was even more consistent within Thaumatomonas enabling CBCs and hemiCBCs (hCBC) to be readily detected by comparing helical regions in closely related strains; these are summarised in Page 15 of 92 16 supplementary Figures 1 and 2 together with the nucleotide differences in 18S rDNA and both ITS1 and ITS2. We found significant evolutionary change between some isolates resulting in CBCs, which could signify an altered function in the processing of the ribosomal RNA precursor. In the case of the Z/S clade, even though NC02 and NC03 t differ from Ts zhukovi and each other by one CBC we decided to rank them only as us cr ip different subspecies of Ts zhukovi because all three have the same 18S rDNA sequence and effectively indistinguishable scales. Another Ts zhukovi isolate HANTSF8 was also made a subspecies because of two hCBCs and distinct scale morphology. Interestingly, the type strain of Ts seravini, with four 18S rDNA differences from the Ts zhukovi type strain, had just one CBC compared with the Ts zhukovi type strain, whereas Ts seravini differs from the new Ts zhukovi subspecies by at least 2 CBCs; the number of hCBCs an was consistently higher, making them useful additional species delimitation markers within the Z/S clade. The hCBC is a single‐sided change, which is thermodynamically M stable in RNA molecules and retains base‐pairing; though implied to be half a CBC, hCBC frequency and distribution on the molecule of both CBCs and hCBCs can be ed independent (Caisová et al. 2011). The ITS2 rDNA secondary structure fold for strain 4e (Ts seravini varisquama) was slightly different for helix III compared to the other isolates in its clade; part of the pt stem had an enlarged loop in the same place where CBCs were found in other closely related isolates (Fig.3). Therefore, the enlarged loop could account for the lower than ce expected CBC difference between 4e and two Ts zhukovi subspecies (carolinensis and Ac paracarolinensis). Checking the folds by eye helped find this difference. Setting boundaries between closely related Thaumatomonas species Figure 2 shows that the zhukovi/seravini subclade has marked intrastrain variation in scale structure. We did not confirm that the type strain (T‐3) of Ts zhukovi had all the scale types, triangular‐like and oval, observed in the original description by Mylnikov and Mylnikov (2003); when examined in the TEM we saw only two cells, both with triangular scales only (Sup. Fig. 6). Seven strains in the Z/S clade were polymorphic for oval and triangular or quasi‐triangular scales, whereas only two have just oval and one just triangular scales. Therefore in the Z/S clade one cannot use scale Page 16 of 92 shape alone to identify species; morphology must be combined with sequences. Yet 17 18S rDNA sequences showed low divergence amongst these closely related Thaumatomonas; Ts seravini and Ts zhukovi differ by just four 18S rDNA nucleotides. In the Ts coloniensis/oxoniensis subclade, we found only one strain polymorphic for t oval/triangular scales. Though scale structure is more uniform in this subclade low 18S us cr ip rDNA divergence makes species discrimination difficult. It is unwise to make separate species just because strains differ in 18S rDNA by one or two nucleotides, unless one thoroughly excludes sequencing error and shows that ITS rDNA sequences are even more different. In addition to comparing ITS2 rDNA secondary structures among closely related strains to use CBCs and hCBCs as an aid to deciding species boundaries, we examined our larger 18S rDNA alignment with many more closely related an Thaumatomonas isolates to seek likely sequencing errors (in new and published sequences) to reduce such errors in judgment. Overall we decided that our numerous M new strains for these two Thaumatomonas subclades are all best included in four existing species and a fifth ‘species’ (Ts vancouveri) should now be included in Ts ed coloniensis as a subjective junior synonym: The Ts coloniensis/oxoniensis subclade (predominantly oval scales) and suppression of Ts vancouveri. We isolated two strains effectively indistinguishable pt from Ts coloniensis (PML3A, Sup. Fig. 3GH, and MLTB12, Sup. Fig. 4D), and two other cultures (GMPL1, Sup. Fig. 3A – C, and EP3, Sup. Fig. 3D ‐ F) similarly indistinguishable ce from the type strain of Ts oxoniensis. These were not included in the 18S rDNA tree because the 18S rDNA sequences were identical to type strains. ITS2 shows at least 1 Ac CBC and six hCBCs between these two groups (Sup. Fig. 2). The many new Thaumatomonas sequences revealed likely errors in published sequences, including Ts vancouveri (AF411264) (Howe et al. 2011a), which appears to have missing nucleotides and sequencing errors. Given its likely sequence errors, Ts vancouveri probably really has the same sequence as Ts coloniensis (both strains HFCC59 type, and HFCC93, which have identical 18S, and, apart from 3 single nucleotide indels that might be sequencing errors, identical 28S rDNA) except for one inserted C in Ts coloniensis, which could also be a sequencing error. ITS2 rDNA sequences are unavailable for the type strain of Ts coloniensis (HFC59 or for other original Ts coloniensis strain HFCC93), but ITS1 of Page 17 of 92 18 HFCC93 was included in Figure 2 to represent the species as its branch was shorter on an ITS1 tree than was HFC59 whose ITS1 differs by two substitutions and three 1‐2 nucleotide indels. The ITS1 rDNA sequence of Ts ‘vancouveri’ is almost identical to our PML3A Ts coloniensis strain and both branch so closely to HFCC93 Ts coloniensis (closer t than does our second coloniensis strain, MLTB12) that we cannot treat them as us cr ip separate species. The above‐mentioned ‘inserted C’, in the 18S rDNA sequence, is the only definite difference between Ts coloniensis and Ts oxoniensis (HQ121429) but Ts oxoniensis has 77 undetermined nucleotides (n) in its sequence (Howe et al. 2011a) (Sup. Fig. 2). Ts vancouveri and Ts oxoniensis were made separate species by Howe et al. (2011a) using ITS2 folding analyses; two CBCs were found, making them likely different species (Coleman 2009, Wolf et al. 2013). Despite Ts coloniensis not having an ITS2 rDNA sequence available for secondary structure analysis, in our new ITS2 rDNA CBC analysis the cluster including Ts vancouveri and our two new Ts coloniensis strains M on the one hand and all three strains of Ts oxoniensis on the other consistently shows the ITS2 rDNA difference previously used by Howe et al. (2011a) to separate Ts vancouveri and oxoniensis, which is congruent with molecular phylogenies (Fig. 2). ed However, as our new Ts coloniensis strains reveal that there is no clear ITS2 distinction between Ts vancouveri and coloniensis, we now make Ts vancouveri a synonym for Ts pt coloniensis but keep Ts oxoniensis as a distinct species because the original CBC distinction of Howe et al. (2011a) is confirmed in three independently isolated strains ce for each species (in disagreement with Karpov (2011), who made Ts oxoniensis a synonym of coloniensis and considered Ts vancouveri invalid because of incomplete Ac scale description). The Ts zhukovi/seravini clade (predominantly triangular scales). We sequenced 18S rDNA sequences from eight new strains of Ts zhukovi (T‐3, UPL1Be2, UPL1Bf2, CCL4B, GMKL4, NC02, NC03, HANTSF8), of which three were made subspecies determined by ITS2 rDNA secondary structure: HANTSF8, NC02 and NC03. A ninth isolate, GMBGL1, with no 18S rDNA sequence, is also within this clade. Ts zhukovi type, T‐3, had at least 12 nucleotide differences in ITS1 rDNA from other Ts zhukovi strains, and at least seven in ITS2 rDNA. However, ITS2 rDNA secondary structure yields no CBCs or hCBCs between Ts zhukovi strain T‐3 and five other Page 18 of 92 isolates: UPL1be2, GMBGL1, UPL1Bf2, CCL4B and GMKL4. Compared with T‐3, 19 HANTSF8 had no CBCs but 2 hCBCs, whereas NC02 had three CBCs and four hCBCs, and NC03 had two CBCs and 3 hCBCs, which made NC02 most divergent of the nine strains for ITS2 rDNA folding. ITS2 rDNA sequences for all nine isolates differed by at least one t nucleotide, except for GMBGL1 and CCL4B, which have identical ITS2 (supplementary us cr ip Figure 1). There were three 18S rDNA sequences for new Ts seravini strains: T‐2, 4e, and CH8; ITS rDNA regions were sequenced only for T‐2 and 4, maximally supported as sisters (Fig. 2). an Sorting out the Thaumatomonas sequence labelling confusion Howe et al. (2011a) concluded from contradictory labelling of essentially the same Thaumatomonas sequences in GenBank, that labelling of either the M Thaumatomonas genes sequenced by Ekelund et al. (2004) or the three source cultures originating from Mylnikov via Karpov and then Vickerman had been mixed up, ed since one of them was ostensibly the same T‐2 type strain Ts seravini culture sequenced by Cavalier‐Smith and Chao (2003) (Mylnikov gave them directly in 1994 in St Petersburg), but the sequences were contradictory). Ekelund (pers. comm.) tells us pt that his strain labels were independent of those of Mylnikov and Karpov (1993); but he coincidentally used virtually the same strain codes T1‐T3 for different species: SCCAP ce T1’ Ts sp., ‘SCCAP T2’ Ts lauterborni, and ‘SCCAP T3’ Ts seravini. By contrast Mylnikov’s T‐1 is Ts lauterborni, his T‐2 is the Ts seravini species type strain, and T‐3 is the Ts Ac zhukovi species type strain. Howe et al. (2011a) incorrectly assumed that the Ekelund labels without hyphens came from those of Mylnikov with hyphens; yet that contradictory use of essentially identical numbering does not explain all contradictions discovered by Howe et al. (2011a). We resequenced and reexamined scale ultrastructure of a fresh sample of the original Ts seravini (T‐2) strain from Mylnikov directly given in 2010 by Denis Tikhonenkov to TC‐S. Its sequence is identical to Ts seravini AF411259 of Cavalier‐Smith and Chao (2003) and also ‘Ts lauterborni SCCAP T2 AY496045’ of Ekelund et al. (2004), but very different from AY496044 of Ekelund et al. (2004), clearly mistakenly labelled ‘SCCAP T 3 Ts seravini’. Figure 4G shows that this Page 19 of 92 fresh T‐2 Ts seravini culture has uniformly triangular scales indistinguishable from 20 those originally described for this type strain by Mylnikov and Karpov (1993) and also by Mylnikov and Mylnikov (2003). Therefore, neither Mylnikov’s nor our laboratory mixed up this culture with another; by contrast AY496044 and AY496045 certainly t have the wrong names in Ekelund et al. (2004) and GenBank, exactly as Howe et al. us cr ip (2011a) indirectly deduced. Furthermore, two strains of new species Thaumatomonas constricta have very similar scales to Ts lauterborni of Mylnikov and Karpov (1993), but quite distinct from those of the other three species: oval scale with no upper tier perforations (Fig. 5B & F). Our sequences for these are closest to AY496044 (Ts seravini SCCAP T3) of Ekelund et an al. (2004), giving independent evidence that ‘Ts seravini SCCAP T3’ almost certainly came from Ts lauterborni (T‐1 of Mylnikov). Therefore, at some stage the names Ts seravini and Ts lauterborni for these two cultures must have been swapped during M travels through four laboratories. Discovering the correct species name (lauterborni) for AY496044, together with Ekelund’s labelling information, invalidates the ed assumption (Howe et al. 2011a) that Ekelund’s T1 was Mylnikov’s T‐1. pt Thaumatomonas zhukovi and Vickerman’s Thaumatomonas sp. An authentic sample of Ts zhukovi type strain (T‐3) (Mylnikov and Mylnikov 2003) was given to TC‐S directly ce by Denis Tikhonenkov. Its 18S rDNA sequence is distinct from all previously published, though differs from Ts seravini (T‐2) by only four distinct nucleotides (top matrix Sup. Ac Fig. 1). Therefore no previously published sequences under other names could mistakenly have come from Ts zhukovi (T‐3). However, only triangular scales were observed in our single TEM preparation (two cells viewed) for Ts zhukovi (T‐3) (Sup. Fig. 6D); no oval scales like those also pictured in the type description were seen. Tikhonenkov donated for our free use his unpublished partial 18S rDNA sequence of Ts zhukovi T‐3; where overlapping with ours it was identical; KC243119 is a long contig including his and ours. All three Russian Thaumatomonas strains isolated by Mylnikov now have reliably identified 18S rDNA sequences: Ts lauterborni (T‐1) dead (AY496044); type strain of Ts seravini (T‐2) (ATCC 50636) (AF41125, EF455776, Page 20 of 92 21 AY496045); and type Ts zhukovi (T‐3; CCAP 1974/4; KC243119). All three Thaumatomonas sequenced by Ekelund et al. 2004 (two wrongly labelled) were the same as those sequenced earlier by Cavalier‐Smith and Chao (1995, 2003), one (Ts seravini) sequenced a third time by Yoon et al. (2008). Figures 1 and 2 have the us cr ip t corrected species names for the Ekelund et al. (2004) sequences. Ekelund’s Ts sp. sequence (AY496046) closely matches sequence U42446 Thaumatomonas sp., one of Vickerman’s cultures published by Cavalier‐Smith and Chao (1995). Howe et al. (2011a) cautiously labelled sequence AY496046 ?lauterborni, which we now know was wrong. From Ekelund’s new information, it came from a soil strain isolated by Vickerman (originally labelled Allas sp.) and donated to Ekelund; an Ekelund (pers. comm.) changed the name to Thaumatomonas sp. ‘because they were then synonyms’ (but on that view the older name Allas would be the correct one for his Thaumatomonas sequences: see Howe et al. 2011a). To end past confusion conformity with our conclusions. M GenBank sequence names from Ekelund et al. (2004) should be corrected in ed These additional Thaumatomonas sequences make it clear that U42446 has sequencing errors, through manual sequencing using autoradiography, without which pt AY496046 matches it completely. Evidently Ekelund’s Ts sp. from Vickerman was exactly the same strain as Vickerman’s Ts sp. earlier sequenced by Cavalier‐Smith and ce Chao (1995). Vickerman recently gave us a culture from his collection labelled ‘Thaumatomonas lauterborni (Karpov)’, stating that “On revival I am not sure what it is, Ac as it does not seem to behave like the original description… So Karpov’s isolate may have got mixed up with someone else’s isolate (e.g. mine).” We found its 18S rDNA sequence to be the same as ‘Ts sp.’ AY490646; its scales (Sup. Fig. 5) are most similar to Ts coloniensis. Probably Vickerman incorrectly relabelled Ts sp. strain ‘Ts lauterborni Karpov’ after giving it (correctly labelled Allas/Th sp.) independently to Cavalier‐Smith and Chao (1995) yielding U42446, and to Ekelund et al. (2004), yielding AY490646. The automated sequence AY490646 is more reliable than U42446 for Vickerman’s Ts sp. strain, which we now designate Ts coloniensis. Confusingly, Howe et al. (2011a) said U42446 has scales ‘similar to Ts lauterborni’, based on Vickerman’s early observations when the only other known species was Ts seravini with triangular scales (Ts zhukovi Page 21 of 92 and coloniensis were still undescribed); Vickerman’s remark meant only that Ts sp. 22 scales were oval with perforations in the upper tier rather than triangular, not that it was closer in detail to lauterborni than subsequently described species with broadly similar perforated oval scales. These sequencing errors mean that some fine detail of t the Ts coloniensis part of the Figure 1.1 tree must be dominated by sequence errors, us cr ip not real differences; Figure 2 more reliably represents diversity and relationships in that half of the Thaumatomonas clade. Taxonomic revisions, new species, and microscopic observations As both phylogenetic trees indicate (Fig. 1 and Fig. 2), scale structure maps an coherently onto phylogenetic trees and so is given special taxonomic emphasis. Splitting Thaumatomastix into three genera. All Thaumatomastix as redefined by Howe et al. (2011a) are marine or freshwater thaumatomonads with spine scales as M well as triangular or oval plate‐scales. Their long spines are of two radically different kinds, with radially symmetric bases in marine species, but asymmetric ones in ed freshwater species (Nicholls 2012b), giving three separate combinations of distinctly different scale types. Our trees showed that each of the two marine clade has either triangular or plate scales, never both, but both have similar radially symmetric spine pt scales; as these clades are not sisters, morphology and plate scale ultrastructure congruently show that they should be made separate genera: Thaumatospina for the ce deepest branching clade with triangular scales and long scale‐free anterior cilium; Ovaloplaca with oval scales and, unlike all other thaumatomonads and all other Ac Cercozoa, a very short anterior cilium clothed in tiny scales. Ovaloplaca is sister to another new marine genus (Scutellomonas) that like it has oval body plate‐scales, but lacks spine scales, so cannot be included in Thaumatomastix. Unfortunately, ultrastructure is unknown for the type species Thaumatomastix setifera (Lauterborn 1899). However, comparison of the original drawings of Lauterborn with the light micrographs of Nicholls (2012b) of the only other freshwater species, Tx triangulata and Tx nigeriensis, reveals a previously overlooked feature of cell structure shared by all three freshwater species and absent from all marine ones studied by light microscopy. This makes us reasonably confident that all three Page 22 of 92 23 freshwater species should be kept in the same genus, so we restrict Thaumatomastix to them alone. These three freshwater species uniformly differ from all marine ones for which rDNA sequences are available in having a conspicuous 1 μm thick cell boundary layer visible in the light microscope as a ‘boxed’ look to the edge of the cell. t This feature is evident from Nicholls’ (2012b) ultrastructural studies of Tx triangulata us cr ip and Tx nigeriensis; the boxed appearance is created by the plate scales’ two‐tier structure, in which the scales are about 1 μm thick and the two tiers are well separated by broad columns. This wide separation of the two tiers makes the ordered boundary layer of the cell more conspicuous than in marine species; on Lauterborn’s (1899) drawing it is 1 μm thick. All marine species so far described except Tx tripus (Takahashi and Hara 1984), which as discussed below might be in yet another genus, an have thinner scales only about 0.5 μm thick (Nicholls 2012b; Thomsen et al. 1993); though most have not been seen as living cells, in two species of Thaumatospina M studied by us and one by Chantangsi et al. (2010) and in our new Ovaloplaca the scale‐ containing boundary layer is significantly less thick (~0.6 μm) and thus distinctly less conspicuous. We therefore conclude that the plate scales of freshwater Tx setifera are ed about 1 μm thick, making it unlikely to belong in the same genus as any of the sequenced marine species, given the deep phylogenetic separation on our trees of pt thaumatomonads with substantially different scale types. Accordingly, we here restrict Thaumatomastix to freshwater species with 1 μm ce thick plate scales, now excluding the marine species, which have scales twice as thin. As explained in more detail below, spine scales of the freshwater Tx triangulata and Tx Ac nigeriensis are radically different from, and probably not homologous with, those of marine former Thaumatomastix; those of Tx setifera are more likely than not to be similar to these two freshwater species. The marine ‘Thaumatomastix’ of Larsen and Patterson (1990) from Wailupe Beach, Hawaii, was not in our view reliably identified as the same species as freshwater Tx setifera. It was the same length and the illustrated cell a bit wider and otherwise indistinguishable from the Thaumatomastix sp. micrograph of Chantangsi et al. (2010) that our trees show is an undescribed Thaumatospina (if its scale layer was drawn to scale, it was only about 0.5 μm thick). Thus we are unaware of any convincing marine records of Tx setifera. We predict that Page 23 of 92 24 genuinely freshwater Tx setifera will turn out to have triangular plate scales and non‐ bobbin‐based radially asymmetric spine scales like both other freshwater species, and frame the revised diagnosis on that assumption: Thaumatomastix Lauterborn, 1899. Revised diagnosis: Freshwater non‐ us cr ip t plasmodial thaumatomonads with prominent nucleus, spine scales, and triangular 1 μm thick plate scales visible in light microscope as a peripheral double layer with ‘boxed’ substructure. Plate scales two‐tiered with upper and lower tier of the same size. Spine scales like somewhat smaller plate scales with long eccentrically placed spine, without basal flange, stemming at a slant from beside central hole in two‐tiered base. Plate scales and spine‐scale triangular bases with gently rounded vertices. an Posterior cilium longer than cell; non‐scaly anterior cilium half to 2/3 cell length. Ventral pseudopodia. Type species: Thaumatomastix setifera Lauterborn, 1899. Other Two types of spine scale M species: Tx triangulata (Balonov 1980), nigeriensis (Wujek 2008). ed Both marine former Thaumatomastix clades have spine scales that are very different from freshwater Thaumatomastix, but closely similar to each other, with a long spine stemming from a base that is several times smaller than a plate scale and pt bearing a circular flange a bit higher up that is narrower than the base plate. The base plate, flange, and the connector joining them are radially symmetric, but the ce connector flares out towards the base plate (Nicholls 2012b; Thomsen et al. 1993), this flared region becoming three physically distinct splayed out struts in Ovaloplaca alone. Ac The flange is narrower than the base plate; apart from that, the overall appearance of spine scale bases of marine ‘Thaumatomastix’ is like a radially symmetric bobbin. Therefore we refer to these two structurally related but distinct marine spine scales as bobbin‐based, to contrast them with the radically different spine scales of two freshwater species: Tx triangulata and Tx nigeriensis (Nicholls 2012b). Though not sufficiently emphasising their extremely different structure from marine spines, Nicholls (2012b) showed that both freshwater species have eccentric spine scales that consist of two closely associated structures: the spine itself (with no sign whatever of a radially symmetric bobbin‐like base or flange) is eccentrically placed on a very broad Page 24 of 92 25 base that closely resembles the plate‐scales of the same species in being composed of two tiers of equal diameter. This two‐tier spine scale‐base has a somewhat smaller diameter than the standard spineless plate‐scales in the same species, but with similar appearance and separation (~1 um) between the two tiers. The spine shaft is plain t with a tapered tip, its extreme base having a subtle flare, and joined to a plate‐scale‐ us cr ip like base, not centrally and vertically as in bobbin‐based scales, but to one side of its central hole and slanting at an angle. We consider the spine scales of both marine clades as mutually homologous, so deduce that they originated in the last common ancestor of Thaumatomonadidae and were lost independently by Scutellomonas and the Reckertia/Thaumatomonas/Allas an clade. We cannot regard the radically different spine scales of the freshwater species (Thaumatomastix sensu stricto) as directly homologous with the bobbin‐type scales of the marine species (see discussion for detailed arguments about their homology and M evolution). New Genus Thaumatospina. Diagnosis: Marine thaumatomonads producing ed plasmodia. Bobbin‐like two‐tiered triangular plate‐scales up to 0.5 μm thick, seen as a 0.6 μm thick boxed outer edge to cell by light microscopy; plate‐scale tiers of equal pt size, vertices rounded, differ in internal structure. Upper tier dish‐like with broad margin and steep‐sided smaller central triangular depression, whose sides appear in ce surface view as thin dense triangular line about 2 μm inside scale edge (in side view the dish floor and side junction is sharp right angle), unlike most Reckertia usually Ac without corner holes (apparent exceptions Ta patelliformis and dybsoeana); upper tier connected to flatter lower tier by a strut just inside each corner of central depression, seen in surface view as elongate dense spot just inside vertex of inner triangular dense line. Long spines (visible in LM) extend centrally from radially‐symmetric bobbin‐like spine‐scale bases with a flange some distance above base‐plate; flange markedly narrower than basal disc; connector between flange and base plate distally and basally flared, often with triple‐strutted substructure within basal flare; basal disc somewhat triangular with rounded vertices, small enough to fit snugly into central depression of plate scale; shaft between basal disc and flanged ridge and flared at base. Anterior Page 25 of 92 cilium obvious, third to half cell length; cilium scales not reported. Type species 26 Thaumatospina vancouveri sp. n. Etym: Thaumato Gk wondrous spina L. spine. Type strains, illustrations, and sequences (or any two) in all species diagnoses below are to be regarded as part of a ‘syntype’, which comprises all or some of the us cr ip t following: living culture, illustrations, and sequence(s) as explained in Cavalier‐Smith and Chao (2010). Cell measurements were from gliding single cells only (not plasmodia) for cell length (CL) and width (CW) at widest point; their ratio (ovalness ratio OR) objectively indicates approximate shape numerically (1.0 round; > 1.0 elongate; <1 wide). Posterior cilium (PC) length is an average; it is also shown as X CL ratio for ready comparison with average CL. Scale dimensions are averages with ranges an in parentheses: scale length (SL) and scale width (SW). All scales are two‐tiered and often have a smaller top tier in Thaumatomonas and Allas. The bottom tier (bt) is the primary measurement given in diagnoses; oval scale averages lengths (oSL), width M (oSW) and OR are given, with ranges in parentheses. Where possible, top tier measurements are given (also in Table 1). The Thaumatomonas and Allas scale’s ed middle upper portion is joined to the perimeter of the upper tier by delicate perpendicular struts associated with holes or perforations on sides, which are not always distinct and often broken. The number of these lateral irregular‐shaped holes pt given as a range refers to one side only. ce We describe three new species with over ten 18S rDNA nucleotide differences and contrasting cell shape/growth forms, but very similar scales, most like those of Ta patelliformis except for spine scales having three unflared terminal teeth, not two flattened flared‐out ones Ac as in patelliformis (Takahashi and Hara 1984): Thaumatospina vancouveri sp. n. Type Figure 6A‐B. Diagnosis: CL 11.9 µm (8.6 ‐ 18.6; N = 39); CW 6.7 µm (4.6 ‐ 9.1); OR 1.8 (1.1 ‐ 3); PC 17.7 µm (15 ‐ 20.9); 1.6 X CL (1.1 ‐ 2.4). Cells strongly elongate, bulbous; ventral groove up to 2/3 CL. AC, spines, and thin (0.6 µm) ‘boxed’ plate scale layer visible in LM (Fig. 6). PC tip sometimes has flaring bifurcation. Ventral finger‐ like branching pseudopodia, tips pointed. Larger polynucleated masses or aggregates common. Strongly curved tapering hollow spine‐scale; tip slightly flared, open with three points. Spines highly curved, length 4.0 µm (3.3 – 5.3; N = 17), tips narrowly trifurcate, unflared; base plate more obviously triangular than in Ta mexica and arabica, width 0.64 µm (0.5 – 0.79); flange Page 26 of 92 about half base plate width. 0.28 µm (0.19 – 0.40) from base plate. Two‐tier rounded 27 triangular plate‐scales with inner electron‐dense line of same shape; each side 1.2 µm (0.76 – 1.50; N = 20); electron‐dense strut somewhat curved, thicker at inner edges of inner electron‐ dense line; central region varies, some scales (especially smaller ones) with lighter inner triangle with vertices opposite centre of flat sides (Fig. 6B arrows), some large ones with us cr ip t similar central light triangular area with vertices opposite scale vertices, some with additional smaller central paler rough and indented/lobed circular of various widths; upper tier has punctate markings at outer margin (barely visible in Figure 6B), apparently small tooth‐like projections from its underside. Type strain: VWB1C (From marine surf, Wreck Beach, Vancouver, Canada. JMS). Type 18S rDNA sequences GenBank KC243114. Etym: after Vancouver city. Comment: As cell size was so variable several clonal cultures were made from single picked cells, but size variability remained. The offset light triangle on some plate scales an might be a print left by a spine scale basal disc attached to it in vivo, which could explain why only some scales show it; however, the fact that Ta arabica has a similar offset pale triangle M but circular base plates argues against this. Thaumatospina mexica sp. n. Type Figure 6C‐F. Diagnosis: Round to oval, bulbous, AC in constant motion when gliding. Finger‐like pseudopodia. Cell size 6.5 ‐ 8.4 µm long in TEM, ed shrunken (no live cells measured). Multiciliate and polynucleate gliding cells common. Spine‐ scales visible in LM. Spine‐scales and triangular body‐scales as for Ta vancouveri except body‐ pt scales consistently have roughly lobed central depression with larger, lighter rounded area overlain on the central inner electron‐dense rounded triangle (vertices never opposite flat side ce as in vancouveri). Spine‐scales curved to almost straight, length 4.1 µm (3.1 – 5.6; N = 18); base plate much less triangular than Ta vancouveri, slightly less circular than Ta mexica, width 0.58 µm (0.5 – 0.64); flange height from base plate 0.28 µm (0.21 – 0.40). Triangular scales 1.2 µm Ac (0.93 – 1.43; N = 28). Type strain: TxMex (Marine, Mexico. JMS). 18S/ITS1&2 rDNA sequences GenBank KC243115/KF577833. Etym: Mexica from Mexico. Comment: 18S rDNA has a large 400 bp insert (position 1158 ‐ 1558) (not repeated elsewhere) that does not BLAST to any GenBank sequence. Repeated PCR with different primers yielded the same insert. Thaumatospina arabica sp. n. Type Figure 7B ‐ C. Diagnosis: larger than Ta vancouveri, CL 14.9 µm (12.7 ‐ 18.6; N = 13); CW 10 µm (9.1 ‐ 11.8); OR 1.5 (1.2 ‐ 1.9); PC 20.8 µm (20 ‐ 22.7); 1.4 X CL (1.1 ‐ 1.6). Oval to round and bulbous, long AC. Glides slowly and jiggly. Spine‐ scales obvious in LM. Refractile granules common in cell posterior. Lamellar and finger‐like pseudopodia, possibly reticulose, sometimes produced ventrally when sessile. Ventral groove up to 2/3 length of cell. Spine‐scales as in Ta mexica but flange edge more rounded than in Ta Page 27 of 92 28 mexica and vancouveri (no measurements as scale bar absent from micrographs); spine hollow, flares out below flange as three struts onto near‐circular base plate, tapers incompletely to subtle, short sharp trifurcate tip, two outer points more prominent than third. Two‐tiered triangular plate‐scales with rounded vertices, distinct inner margin, obvious evenly spaced punctate underside margin; prominent inner line with inner three‐pronged mark inside each us cr ip t corner; central deep triangular depression shows light central triangular area with 60° offset as in some vancouveri scales. Type strain: ThxDubai. (Marine surf, Dubai. JMS). Type 18S/ITS1&2 rDNA sequences GenBank KC243113/KF577834. Etym: arabica L. from Arabia. Comment: The cell in Fig. 12 of Takahashi and Hara (1984) had spine tips more like Ta arabica than the type Ta patelliformis (their fig. 9), so might be one of the three species described here not Ta patelliformis. Thomsen et al. (1993) found another example of similar spines. Takahashi and Hara (1984) said there was no spine variation on the same cell but that on others spines are an straighter with slightly flattened bifurcate tips. Possibly Ta patelliformis in these earlier papers included more than one species. ‘Thaumatomastix’ sp. of Chantangsi et al. (2010), robustly M sister to but genetically distant from our three species, is a fourth (undescribed) Thaumatospina with slightly longer spines. New combinations. We transfer six former marine Thaumatomastix with similar scales ed to Thaumatospina. They fall into two contrasting morphological groups; plate scales of (b) are so different from those studied here that they might eventually merit recognition as a distinct pt subgenus (or genus if sequencing supported that): ce (a) Three species with spines and plate scales similar to Ta vancouveri Thaumatospina patelliformis comb. n. Basionym Chrysosphaerella patelliformis Takahashi and Hara (1984 p. 106). Synonym Thaumatomastix patelliformis (Takahashi & Hara) Beech and Ac Moestrup, 1986 Thaumatospina curvata comb. n. Basionym Thaumatomastix curvata Nicholls (2012b) Thaumatospina sablensis comb. n. Basionym Thaumatomastix sablensis Nicholls (2012b) (The figure show that its diagnosis statement that spine‐scales are 40 to 52 μm long is obviously a typo for 4.0‐5.2.) (b) Three species with substantially larger plate scales (struts nearer centre than edge) than Ta vancouveri and shorter, thicker less curved spines Thaumatospina tauryanini comb. n. Basionym Thaumatomastix tauryanini Mikrjukov (2002) Page 28 of 92 Thaumatospina inornata comb. n. Basionym Thaumatomastix inornata Nicholls (2012b) 29 Thaumatospina gwaii comb. n. Basionym Thaumatomastix gwaii Nicholls (2012b) Thaumatomastix tripus (Takahashi & Hara) Beech and Moestrup, 1986 differs from Thaumatospina as here defined in having its plate scales with 1 µm separation between the us cr ip t tiers; it differs from freshwater Thaumatomastix as here redefined by having radially symmetric spine scales with bobbin‐like bases, and much longer spines than either Thaumatospina or Thaumatomastix. It therefore does not fit into any genus as here defined; until we have sequences for T. tripus and/or Tx setifera we leave it in Tx, even though it may not belong there; like Beech and Moestrup (1986) and Nicholls (2012b) we now consider T. tripus a thaumatomonad (not Chrysosphaerella) and that the original mention of a chloroplast, and perhaps also contractile vacuole (unusual in marine species), were errors. It might belong an to the uncultured marine lineage branching between Thaumatospina and Ovaloplaca/Scutellomonas; its plate‐scale central region is obviously different from the three M Thaumatospina we sequenced and from the other six species by having three radiating lines from the centre as in Tx triangulata (Nicholls 2012b). New Genus Ovaloplaca. Diagnosis: Marine non‐plasmodial thaumatomonads; ed light microscopy shows inconspicuous 0.6 μm plate‐scale layer and conspicuous spines. Two‐tiered oval plate‐scales up to 0.5 μm thick, both tiers of similar size, upper rim pt thicker, or upper tier somewhat (up to 15%) smaller; often perforated, often perforated, perforation pattern differing between species. Long spines extend ce centrally from radially symmetric bobbin‐like spine‐scale bases less than half the width of a plate‐scale, with flange above base‐plate; flange markedly narrower than base Ac plate; connector between flange and base plate distally and basally flared, with triple‐ strutted substructure resembling base of Eiffel tower within basal flare. Anterior cilium small and inconspicuous, sometimes bearing much smaller round to oval single‐tier scales. Ventral branching, long, slender pseudopodia. Type species: Ovaloplaca salina (Birch‐Andersen) comb. n. Etym: Ovalis L. oval, placa Galician plate, referring to oval plate‐scales. We describe one new species and make four new combinations: Ovaloplaca yabukii sp. n. Type Figure 8E‐G. Diagnosis: CL 11.1 µm (N = 1); CW 10.5 µm (x ‐ x); OR 1.0; PC 11 µm; 1X CL; AC not noticed. Cells glide seldom, usually gyrate in one place. Spine scale 5.18 µm (4.5 ‐ 6.2; N = 9) with longitudinally triple ribbed spine with shortly Page 29 of 92 trifurcate tip, slightly contorted but not completely twisted; straight or strongly curved near 30 tip; prominent basal flange splays out as three short, broad struts onto rounded‐triangular or round base (base ~0.5 µm diameter). Oval plate scales; tiers same size; prominent two‐level central transverse band separates two D‐shaped bowls; bowls regularly perforated all round edges, rarely (as O. salina typically is extensively) with very few in centre. oSL x oSW; 1.29 x us cr ip t 0.73 µm (1.2 ‐ 1.4 x 0.6 ‐ 0.9; N = 6); OR 1.8 (1.4 ‐ 2.2). Type strain: YPF609 (TKB‐345 deposited as NIES‐2378. Marine. Tokyo Bay, Japan. A. Yabuki). Type 18S rDNA sequence differs by one nucleotide (C at position 678 instead of T at 715) from O. salina Norwegian strain UIO286 (FR846196) (Ota et al. 2012) in a variable region (also a few undetermined nucleotides): GenBank KC243118. Etym: after Akinori Yabuki (isolator and provider of the type sequence and micrographs). Comment: Very similar to O. salina (Beech and Moestrup 1986), which is more elongate and whose spine scales vary more in length. O. salina has only one definite nt an difference. Because the type strain of O. salina no longer exists we accept UIO286 as representing it, as it is extremely well characterised and like the type was from Scandinavia, M and treat Yabuki’s Japanese strain as a separate extremely closely related species distinguishable by 18S rDNA sequence and plate scale perforation pattern (O. yabukii more often has perforations in the central region of the D‐shaped depressions); O. salina has oval AC ed scales. Takahashi and Hara (1984) noted that Japanese O. salina was more variable in plate‐ scale perforation patterns and spine tip tooth length than the type or other Danish samples. Yabuki’s culture is more uniform and most similar to their Fig. 18, and we did not see either pt the extremes of plate scales with no perforations or perforated all over the D‐shaped regions reported previously. Several similar, subtly distinct cryptic species may coexist in Japanese seas. ce Plate scales of O. salina and yabukii have an unperforated broad central transverse band where upper and lower tiers are not in contact, unlike the uniformly perforated O. multipora Ac or even more different unperforated O. asymmetrica scales. Four new combinations: Ovaloplaca salina comb. n. Basionym Chrysosphaerella salina Birch‐Andersen, 1973. Synonym Thaumatomastix salina (Birch‐Andersen) Beech and Moestrup, 1986. Ovaloplaca bipartita comb. n. Basionym Thaumatomastix bipartita Beech and Moestrup, (1986) Ovaloplaca multipora comb. n. Basionym Thaumatomastix multipora Nicholls (2012b) Ovaloplaca asymmetrica comb. n. Basionym Thaumatomastix asymmetrica Nicholls (2012b) Page 30 of 92 New Genus Scutellomonas. Diagnosis: Colourless, heterotrophic dorso‐ 31 ventrally flattened marine thaumatomonads, sometimes floating; round to oval; trailing posterior cilium emerges from subapical pit at bulbous anterior end; anterior cilium not visible in light microscope, possibly sometimes present as very short stub. t Lobose, branching, finger‐like pseudopodia produced ventrally from central furrow. us cr ip Cell covered in two‐tiered oval, unperforated plate‐scales often with thick central longitudinal electron‐dense line ‐ or two lines closely parallel; upper tier appears similar size as lower tier, joined to upper tier conspicuously by inner electron‐dense oval with struts at each end and at centre of sides where it is somewhat pinched in laterally and less conspicuously by these four struts at intermediate positions. Spine scales absent. Type species Scutellomonas patella sp. n. Etym: Scutella L. saucer or flat an bowl, describing scale shape; monas Gk unit. Scutellomonas patella sp. n. Type Figure 8A‐D. Diagnosis: CL 10.6 µm (7.7 ‐ 12.7; N = M 15); CW 8.5 µm (6.4 ‐ 10); OR 1.3 (1.1 ‐ 1.5); PC 15.2 µm (12.3 ‐ 19.31) 1.4 x CL (1.2 ‐ 1.8). Round to oval, dorso‐ventrally flattened cell, wobbles slowly when gliding; bulbous anterior. Ventral groove along nearly entire cell. Fast moving pseudopodia extend just over 1X CL. ed Refractile granules. Round cysts. Acronematic posterior cilium end often stretches when sticks to substrate during gliding. Oval two‐tier plate‐scale commonly (not always) has central thick pt line, sometimes as two close thin lines; electron‐dense oval inner margin appearing ‘pinched‐ in’ where mid and end struts join lower and upper tiers. Prominent thickened margin on both ce tiers. oSL x oSW; 0.77 x 0.53 µm (0.63 ‐ 0.91 x 0.45 ‐ 0.64; N = 30); OR 1.5 (1.2 ‐ 1.7). Type strain: CA05. (Marine. Monterey Bay, California, USA. JMS). Type 18S/ITS1&2 rDNA sequences GenBank KC243116/KF577835. Etym: patella L. dish. Video: Ac http://dx.doi.org/10.6084/m9.figshare.100962. Comment: PC length was difficult to measure on gliding cells because acronematic tips varied in elasticity and stuck to the substrate during gliding, extending/stretching the cilium. No AC observed in light microscope (a white spot sometimes seen at the head of the groove might be an end‐on view of a ciliary stub), but very short scale‐less cilium observed in TEM, perhaps pre‐divisional because not in all specimens. An 18S rDNA sequence, AY180026 from an oxygen‐depleted marine environment (Stoeck and Epstein 2003) is probably another Scutellomonas as it shares very distinct nucleotide signatures throughout with ours. Scutellomonas is most similar to Thaumatomastix thomseni (Tong 1997), not mentioned by Howe et al. (2011a) but transferred here to Scutellomonas. Tong described her cell as different from a similar one in (Vørs 1992b) (‘unkendt flagellat’ ‐ Page 31 of 92 unknown flagellate fig 6.48.); our scales were slightly smaller than Tong’s (a wider range 32 including smaller ones). Vørs and we did not observe cilium scales, but Tong wrote that even though she did not see a second shorter cilium its presence was implied by small scales like those typically on the AC of other Thaumatomastix, which in hers had inner perforations around the margin. However, as Tong studied mixed protist cultures these tiny scales might us cr ip t have come from another cell. We do not consider our isolate conspecific with ‘Tx’ thomseni because Tong’s specimen had a mucilage sheath unlike ours and its central furrow reached only half way down the cell not all along it. New combination Scutellomonas thomseni comb. n. Basionym Thaumatomastix thomseni Tong (1997); synonym Reckertia thomseni Karpov (2011) ‐ its oval scales mean that it was wrongly put in Reckertia, where they are always triangular. an New Reckertia species: Reckertia gemma Type Figure 8H‐J. Diagnosis: Smooth fast‐ gliding oval to round dorso‐ventrally flattened cell with distinct ciliary pocket apically off centre. CL 11.6 µm (10 ‐ 13.2; N = 14); CW 9.7 µm (7.7 ‐ 11.8); OR 1.2 (1.1 ‐ 1.3); PC 19.1 µm M (18.2 ‐ 20.5) 1.6 x CL (1.5 ‐ 1.8). Posterior portion of cell body raised slightly into water column when gliding. Ventrally branching lamellipodia produced centrally, possibly reticulose. Large highly refractile irregular ball‐like cell masses frequent, with multiple cilia. Clumps of irregular‐ ed shaped cells often glide/jiggle. Body scales with plain triangular base (~ 0.45 µm each side) crowned with extremely delicate, very open upper tier connected to corners of lower tier by pt delicate struts that branch dichotomously distally and proximally; upper tier has three‐fold symmetry, consisting of a thicker triangular raised marginal band, each side bearing three ce more slender forked (Y‐shaped) connectors joined at the centre by the fork stems; distal to this latticework is a denser triangular domed ‘jewel’ with curved edges connected at its corners by three extremely slender dichotomous branching strands on either side of the centre Ac of each marginal band inside the junctions with the six forked ends of the Y‐shaped connectors, forming a ‘crown’ that is a third layer distal to the upper tier. Jewel triangle vertex to vertex 0.17 µm (0.15 ‐ 0.2). Oval scales with thickened margin and central thick line thickly coat AC (Fig. 8I); oSL 0.31 µm x 0.17 µm (0.25 ‐ 0.36 x 0.15 ‐ 0.19). Type strain VWB1Thas CCAP 1950/1. (Marine surf, Wreck Beach. Vancouver, Canada. JMS). Type 18S/ITS1&2 rDNA sequences GenBank KC243117/KF577831. Etym: gemma L. jewel. Comment: Scale struts and connectors often broken (thus hard to measure) but the jewel remains intact. Although pseudopodia are generally extruded from the centre of the ventral side, they sometimes seemed to come from lateral surfaces as in the type Reckertia sagittifera (Conrad 1920) that had an extremely long AC, very different from R. gemma’s which is short (Fig. 8J). Even so, a cell with similar Page 32 of 92 dimensions to Conrad’s observed in EM (Thomsen et al. 1993, figure 3) had ciliary scales 33 similar to R. gemma, but a different type of triangular crowned scale with no apparent jewel. Reckertia dybsoeana (Thomsen et al. 1993, figure 15) has similar scales with a jewel in the crown of the triangular scale like R. gemma, but it is not the same species since its cilium scales are circular or nearly so with a non‐elongated central density, markedly unlike the us cr ip t strongly elongated ovals in R. gemma, and the triangular jewel of R. dybsoeana has straighter sides (Thomsen et al. 1993 figures 16, 17 shows body‐scales of an unnamed Danish species which could have been R. gemma; their Finnish scale (Fig. 14) is probably another undescribed species). Figure 3A of Gooday et al. (2006) called Tx dybsoeana was certainly misidentified (the central three‐fold part symmetric of its upper tier has strongly inward curving edges as in Fig. 14 of Thomsen et al. 1993 (but is proportionately much larger), not straight or slightly outward curving edges as in R. dybsoeana), and is probably an undescribed Reckertia species scale with an apparently two distinct superimposed upper‐tier structures of three‐fold symmetry analogous to but very different in detail from R. gemma or dybsoeana scales, suggestive of an M intermediate morphology between the highly dissected upper tier of R. gemma and dybsoeana and more sheet‐like upper tiers of R. filosa (Howe et al. 2011a) and R. formosa that both lack the third crown layer (see discussion for further comparisons). Their Fig. 3B Thaumatomastix ed sp. is a different undescribed Reckertia with even narrower three‐fold central upper tier region without a gem. We suspect that Fig. 3C of Gooday is also misidentified as R. formosa as its upper tier seems smaller than the lower, with supporting struts nearer its corners than in R. pt formosa where both tiers are equal in size. R. dybsoeana was misspelt in the transfer of Thaumatomastix dybsoeana to Reckertia by Howe et al. (2011a) because the legend of the ce type illustration (Figure 14‐17) misspelt it as dybsoena, which contradicted that in the original diagnosis (Thomsen et al. 1993); the correct name is Reckertia dybsoeana (Thomsen et Ac al.1993). Two new Allas species from soil. Both have quasi‐oval, but bilaterally asymmetric, two‐tier scales with a median lateral strut on one side only connecting the upper and lower tiers, as in Allas aff. diplophysa (Howe et al. 2011a) but unlike Thaumatomonas which has end struts only. Both species are smaller than A. diplophysa and their scale upper tiers have many more and smaller lateral holes (Fig. 9) and are obviously mutually different. 18S rDNAs have 19 nucleotide differences between the three species; we noted two distinct Allas 18S rDNA sequence signatures: TCCAT and GACT (at positions 1395 and 1717 of AF411626). Page 33 of 92 Allas multipora sp. n. Type Figure 9A‐B. Diagnosis: Gliding cells often rounded, smaller on 34 average than A. diplophysa. Prominent AC and ventral groove from apical end. CL 12.2 µm (10 ‐ 13.6; N = 16); CW 9.6 µm (7.3 ‐ 13.2); OR 1.3 (1 ‐ 1.7); PC 13.8 µm (12.3 ‐ 15.9) 1.1 x CL (1 ‐ 1.4). Branching pseudopodia not readily observed, quickly retracted. Two‐tier scales with three struts supporting upper tier; end struts predominantly single and broad (two closely us cr ip t positioned narrow ones in A. aff. diplophysa), narrow at base broadening considerably at upper end, appearing in surface view as dense U shape with terminal granules at each end of trampoline‐like upper tier; lateral strut splayed at each end, giving slightly forked appearance, narrower distally than end struts, in surface view resembling a small dense ‘c’ (rarely missing). Some struts misshapen, with two struts at same end. Up to nine small lateral holes on strut free side; fewer on strutted side as hole beside lateral strut is larger. Oval scales; oSL x oSW 0.61 x 0.4 µm (0.44 ‐ 0.84 x 0.28 ‐ 0.49; N = 28; OR 1.7 (1.2‐2.2). Type strain: GSPB9 CCAP an 1903/3. (molehill soil, Galicia, Spain. K. Vickerman. 2001). Type 18S/ITS1&2 rDNA sequences GenBank KC243107/KF577829. Etym: Multi L. many pora L. pores. Comment: Figure 9B M multiple oblique views make it clear that though the lower tier of Allas scales is essentially elliptical, the upper tier is a laterally squashed triangle with a structurally homologous strut supporting each vertex; degree of triangularity varied, some being more oval than others; the ed most oval scales were measured. Being even smaller than A. aff. diplophysa, it is even less like the type species A. diplophysa (average size 18 µm), and merits a new name. pt Allas media sp. n. Type Figure 9C‐E. Diagnosis: Large polynucleate and smaller gliding cells. Smooth gliding with obvious AC; smaller cells with proportionally longer PC than A. ce diplophysa. Filose pseudopodia retracted quickly and often raised in water‐column. CL 11.4 µm (9.6 ‐ 13.2; N = 20); CW 8.9 µm (7.3 ‐ 10); OR 1.3 (1.2 ‐ 1.6); PC 17.5 µm (14.1 ‐ 19.1) 1.5 x CL (1.3 ‐ 1.6). Scale lower tier oval; upper tier rarely oval with only end struts (e.g. Fig. 9D near Ac bottom right), usually a non‐equilateral triangle with rounded vertices each with one U or C shaped strut; up to 10 very small upper‐tier holes/perforations on longest side; two end holes much smaller than in A. diplophysa or A. multipora, sometimes not much larger than hole beside lateral strut. End struts as in A. multipora but U more sharply curved; end holes large, circular or somewhat D‐shaped. Scales; oSL x oSW; 0.58 x 0.47 µm (0.48 ‐ 0.68 x 0.38 ‐ 0.62; N = 41); OR 1.2 (1‐1.4). Type strain: BKARH9. (Beaver Creek Arizona, USA. K. Vickerman. 2001). Dried riverbed sediment. Type 18S/ITS1&2 rDNA sequences GenBank KC243108/KF577830. Etym: medius L. middle, because number of perforations and shape of scale intermediate between other species. Comment: A. media 18S rDNA sequence differs from A. multipora by more than 15 nucleotide differences, and from A. aff. diplophysa by 6; the MrBayes ITS tree Page 34 of 92 that makes A. aff. diplophysa and media sisters better fits this than ML trees that do not. If 35 that topology is correct Allas scales probably had numerous lateral holes ancestrally, reduced in A. aff. diplophysa, and also only three struts consistent with their ancestral scale having been triangular as argued in the discussion. Two other strains from K. Vickerman had the same 18S rDNA sequence: BKARH11 (dried Arizona riverbed); SH4_B4 (Sourhope, Lancaster, UK), so us cr ip t A. media is widely distributed and common. Two new Thaumatomonas species from soil Thaumatomonas solis. sp. n. Type Figure 5C‐D. Diagnosis: Cells broad, often irregular with flared posterior. Branching ventral pseudopodia. CL 14.2 µm (10.9 ‐ 15.9; N = 5); CW 11 µm (8.6 ‐ 12.7); OR 1.3 (1.2 ‐ 1.5); PC 15.9 µm (N = 2); 1.2 x CL (1 ‐ 1.5 N = 2). Two‐tiered oval scales with plain base and margin. Top tier with three or five struts at each end and a central oblong an plate with 4 ‐ 8 lateral holes on each side; end holes roughly D‐shaped; oSL x oSW; 0.57 x 0.31 µm (0.5 ‐ 0.7 x 0.27 ‐ 0.36; N = 26); OR 1.9 (1.6 ‐ 2.3). Type strain: BZ8 CCAP 1974/3 (soil, Brazil. M JMS). Type 18S rDNA sequence GenBank KC243111. Etym: solum L. soil. Comment: Most cells are dorso‐ventrally flattened, but some are so irregularly shaped to be unstable when gliding. Scales most resemble Ts coloniensis and Ts oxoniensis, but Ts solis on average has more lateral ed holes and often only three end struts; when there are five the two end ones are typically closer together suggesting accidental doubling. The closest sequence from strain ATCC50250 (AF411261) has one nucleotide different at the end, possibly an error in which case it should pt be the same species. Strain CH9 (China, soil) Ts aff. solis with six clear 18S rDNA nucleotide differences (KC243112) died before cellular comparisons could be made so is not made a new ce species as it probably deserved; ITS trees suggest that the Ts solis/ Ts aff. solis lineage is sister to the Ts coloniensis/lauterborni/constricta clade (Fig. 2); PhyloBayes for 18S rDNA shows the Ac same (Fig. 1), so its placement as sister to all other Thaumatomonas plus Allas in the ML and MrBayes 18S rDNA trees is probably a long‐branch artefact. This lineage has three distinct nucleotides absent elsewhere in the Monadofilosa alignment (positions 993, 995, 1082). Thaumatomonas constricta sp. n. Type Figure 5E‐G. Diagnosis: Smoothly gliding elongate, oval cells; posterior sometimes acutely pointed. Long branching pseudopodia readily produced, sometimes broad in larger cell aggregates. CL 12.2 µm (6.8 ‐ 15.9; N = 11); CW 7.4 µm (5 ‐ 9.6); OR 1.7 (1.2 ‐ 2.2); PC 15.5 µm (11.4 ‐ 18.2) 1.3 x CL (0.9 ‐ 2.5). Elongate oval body scales; bottom tier plain oval with thickened margin; Top tier narrower centrally than at curved ends, each supported by three (occasionally four) narrow struts beside large hole; slightly domed centrally; lateral holes absent; oSL x oSW; 0.57 x 0.28 µm (0.44 ‐ 0.72 x 0.22 ‐ Page 35 of 92 0.35; N = 25); OR 2.1 (1.7 ‐ 2.4). Type Strain CH3 (soil, China. JMS. 2009). Type 18S/ITS1&2 36 rDNA sequences GenBank KC243110/KF577826. Etym: constrictus L. compressed ‐ to emphasize that the trampoline‐like upper tier central element is narrower than its denser curved supporting ends. Comment: Another Ts constricta strain from Brazil (BZ1, CCAP 1974/2) with identical (but terminally partially missing) 18S rDNA sequence (KC243109) had the same us cr ip t ITS1 sequence, and three differences in ITS2 rDNA (KF577825) but no CBCs and similar scales (Fig. 5B) except that the upper tier seemed to sag downwards rather than bow upwards. BZ1 has a flared acronematic PC tip (figure 5A), not noted in live type strain CH3, though TEM shows a subtly tapered tip with a lighter swollen membranous region (Fig. 5G), similar to Ts zhukovi strain NC02 (Fig. 10H). The dead Panama strain P106 (sequence KF577827) has >20 differences in ITS2 rDNA from Ts constricta, with two CBCs and seven hCBCs making them likely different species, but not describable without micrographs. The scales look most similar an to Ts lauterborni (Mylnikov and Karpov 1993) but dimensions are different: Ts contricta is thinner. M Transfer of Reckertia hindoni to Thaumatomonas: We regard freshwater ‘Reckertia’ hindoni (Nicholls 2012b) with triangular two‐tier scales as a ed Thaumatomonas for four reasons: (1) no scales were observed on the anterior cilium (anterior ciliary scales occur in both phylogenetically well characterised Reckertia plus the type species R. sagittifera, but in no Thaumatomonas); (2) R. hindoni scales are pt completely different from the triangular crowned open lattice type seen in Reckertia gemma and dybsoeana; (3) their struts are more like those of Thaumatomonas ce triangular scales than the curious tripartite type seen in those Reckertia with sheet‐like upper tiers (e.g. R. filosa: Howe et al. 2011a, or R. formosa) and the conspicuous lateral Ac holes along each side, closely resemble those of all triangular scales of Thaumatomonas, notably Ts seravini, whereas no phylogenetically characterised Reckertia have such an arrangement of lateral holes; (4) All Reckertia are marine. R. hindoni scales are most similar to the perforated triangular scales Ts seravini, but R. hindoni does not have extra‐big corner holes, and the other perforations of Ts seravini are larger and the scale vertices are less gently curved in R. hindoni. Considering the variation in perforation size amongst closely related species (Allas, Fig. 9), such differences do not argue against R. hindoni being a Thaumatomonas so we remove it from Reckertia (see discussion). Thaumatomonas hindoni comb. n. Basionym Reckertia hindoni Nicholls (2012b p. 12) Page 36 of 92 Four new Thaumatomonas subspecies: We isolated eight strains of 37 Thaumatomonas strains with very similar rDNA sequences to Ts zhukovi (Fig. 2). Two of these (NC02/3) form a distinct subclade and differ sufficiently in ITS2 rDNA secondary structure (sup. Fig. 1) that it could be argued that they may be separate t species. However Fig. 2 places them as sister to a clade of five strains that do not differ us cr ip from each other or from Ts zhukovi in ITS2 CBCs or hCBCs, and whose scales do not distinguish them unambiguously from Ts zhukovi. As we would regard these five clones as Ts zhukovi, we decided not to treat to NC02/3 as separate species, but make them new subspecies of zhukovi to recognise their ITS distinctiveness. Their scales are very similar: predominantly oval with two inter‐tier end struts but some are more irregular with a third strut making them non‐equilateral triangles (Figs 10AB, GI). an Thaumatomonas zhukovi originally had a mixture of oval scales, equilateral and irregular triangles and even ones with four struts, but when we re‐examined it we saw M only equilateral triangles (Supplementary Fig. 6D). Strikingly the five new strains with zhukovi‐like ITS2 had different ranges of scale morphology: two (GMKL4, GMBGL1; Fig. 10C,D) had only oval scales with 4 or 5 struts at each end and 5‐8 lateral holes on each ed side, and the large terminal holes were shaped like a curved sausage (as in Ts seravini varisquama), not circular as in all other Ts zhukovi; the other three (UPL1Bf2 CCAP pt 1974/1, UPL1Be2, CCL4B; supplementary Fig. 6) had a mixture of oval and triangular scales, usually not exactly equilateral; their oval scales differ from those of GMKL4, ce GMBGL1 in end strut structure, having a single broad end strut with denser edges, which in UPL1Bf2 only may be partially or completely subdivided into two probably Ac separate struts; lateral holes of UPLBf2 are very similar to those of GMKL4/GMBGL1, but UPL1Be2 and CCL4B have fewer larger lateral holes (4‐5 per side). The eighth new zhukovi‐like strain (HANTSF8) also had a mixture of ovals and non‐equilateral triangles but groups on the tree weakly as sister to two Ts seravini strains, which branch from within the Ts zhukovi‐like strains. Furthermore the new Ts seravini strain (4e) has a mixture of oval and equilateral triangular scales in contrast to the original Ts seravini (T‐2, whose scales were all equilateral triangles (Mylnikov and Mylnikov 2003)). These observations indicate that scale shape is developmentally and evolutionarily quite labile in the zhukovi/seravini clade, but that developmental variability may vary from strain to strain (and may even have changed during long‐term culture of T‐3), and is Page 37 of 92 not a sound basis for making separate species. Thus scale variation in our nine new 38 strains and apparent absence now of oval scales from T‐3 have almost broken down the original morphological distinctions between Ts zhukovi and seravini; it would be not unreasonable to make Ts seravini just a subspecies of zhukovi. However, for now t we conservatively retain it as a separate species because T‐2 differs from all 10 other us cr ip strains studied here in one ITS2 CBC. We treat HANTSF8 as a subspecies (saxoni) of zhukovi and 4e as new subspecies varisquama of Ts seravini; morphologically we can say that in this present usage Ts seravini has either equilateral triangle or oval scales whereas Ts zhukovi has oval and or typically non‐equilateral triangular scales. For clarity we first revise the nominal subspecies diagnosis of Ts zhukovi to contrast it with an the others: Thaumatomonas zhukovi zhukovi Mylnikov in Mylnikov and Mylnikov (2003). Revised diagnosis: CL 6.6 – 13.1 µm; CW 4.5 – 8.2 µm, PC 15 – 19.4 µm, 1.1 – 2.3 x CL. Scales variable: M triangular or a mixture of triangular, oval and irregular: triangular scales equilateral or often irregular with elongated upper‐tier apertures and broadly fused struts sometimes with two fusion points distinctly visible. Oval scales have either distinct struts and small upper‐tier ed perforations with between 3 ‐ 8 upper‐tier perforations on one side, or broadly fused struts and darkened round or elongated upper‐tier apertures with 2 ‐ 8 perforations along each side; pt the terminal hole typically circular, rarely elongated curved sausage‐shape. Type strain clone T‐ 3 (Mylnikov and Mylnikov 2003). Comment: Five new Ts zhukovi strains form a distinct ITS ce rDNA clade because ITS1 of the type T‐3 differs by 24 – 25 nucleotides, but we treat them as the same species because ITS2 rDNA is so similar and UPL1B strains have virtually the same scale morphology as T‐3, though the two GM strains, one from Oxford and the other from Ac Keele University, England, clearly differ (Table 3 gives provenance of Thaumatomonas strains). Thaumatomonas zhukovi saxoni subsp. n. Type figure 4, A‐B. Diagnosis: CL 10.3 µm (7.3 ‐ 12.7; N = 23); CW 7.8 µm (5.9 ‐ 10.9); OR 1.3 (1 ‐ 1.9); PC 11.1 µm (6.8 ‐ 12.7) 1.1 x CL (0.9 ‐ 1.4). Round to elongate cell with ventral groove up to 2/3 length of cell. AC short. Sometimes asymmetric apical notch at ciliary pocket. Finger‐like pseudopodia produced ventrally from centre of stationary cell, sometimes extensive, up to 3x CL. Scales two‐tier: oval and irregular‐ shaped triangular, 2 – 10 lateral upper‐tier holes with circular aperture at either end; end holes circular. Bottom tier; oSL x oSW; 0.63 x 0.3 µm (0.45 ‐ 0.85 x 0.28 ‐ 0.38; N = 36); OR 1.9 (1.4 ‐ 2.3). Type Strain HANTSF8 (Molehill, Longstock, Hampshire. UK. KV). ITS1 & 2 rDNA sequence GenBank KF577808. Differs by two hCBCs in ITS2 rDNA secondary structure from the Page 38 of 92 39 Ts zhukovi zhukovi type T‐3 strain. Etym: saxoni L. as collected from a former Saxon region of the UK. Comment: Scales are very regular but some show a propensity to become triangular: rather like Allas but very different from T‐3. The top tier bows upward, not always visible in TEM. The upper‐tier lateral perforations can be up to two holes more than any seen in Ts zhukovi zhukovi. This subspecies is the most rounded (OR 1.3) of all Ts zhukovi zhukovi strains us cr ip t studied. Thaumatomonas zhukovi carolinensis subsp. n. Type Figure 10G ‐ I. Diagnosis: CL 12.2 µm (8.6 ‐ 15.6; N = 20); CW 6.8 µm (4.6 ‐ 9.1); OR 1.8 (x ‐ x); PC 14.4 µm (11.4 ‐ 16.4) 1.2 x CL (1 ‐ 1.6). Smooth gliding with short AC constantly beating. PC tip noticeably plastic, often momentarily sticking to substrate. Oval and triangular two‐tiered scale often with two broadly fused struts at either end with 3 ‐ 5 lateral upper‐tier holes. Circular to oval upper tier end an apertures. Oval scales more common than triangular; end holes somewhat oval. oSL x oSW; 0.55 x 0.33 µm (0.50 ‐ 0.59 x 0.29 ‐ 0.35; N = 10 x); OR 1.7 (1.5 ‐ 2). Type Strain: NC02. (Soil, near stream, North Carolina, USA. JMS. 2010). ITS1 & 2 rDNA sequence GenBank KF577814. hCBCs. Etym: From Carolina state. M Differs by three ITS2 rDNA CBCs from T‐3 Ts z. zhukovi (Mylnikov and Mylnikov 2003), and four ed Thaumatomonas zhukovi paracarolinensis subsp. n. Type Figure 10A ‐ B. Diagnosis: CL 10.7 µm (7.3 – 13.6; N = 16); CW 6.8 µm (4.1 – 9.1); OR 1.6 (1.1 – 2.5); PC 14 µm (11.4 – 15.9) 1.3 x CL (1.0 – 1.7). Smooth gliding with AC constantly beating. PC end noticeably plastic, often pt momentarily sticking to substrate. Mainly oval, sometimes irregular triangular scales; ovals with two broadly fused struts at each end; end holes usually circular, rarely slightly laterally ce oval; 2 ‐ 5 upper‐tier lateral holes. oSL x oSW; 0.54 x 0.32 µm (0.45 ‐ 0.73 x 0.28 ‐ 0.38; N = 18); OR 1.7 (1.3 ‐ 1.9). Type strain NC03 from moss and lichen, North Carolina, USA. JMS. 2010. 18S Ac rDNA sequence same as Ts z. zhukovi T‐3. Type ITS1 & 2 rDNA sequence GenBank KF577815. ITS2 rDNA has two 2 CBCs and 3 hCBC differences from T‐3 Ts z. zhukovi and one CBC compared with Ts z. carolinensis. Etym: para Gk. beside, as it is sister to T. z. carolinensis. Thaumatomonas seravini varisquama subsp. n. Type Figure 4C,D. Diagnosis: Differs from Ts seravini seravini T‐2 (with identical 18S rDNA) by having oval scales as well as triangular ones. CL 11.9 µm (6.8 ‐ 14.6; N = 18); CW 9.5 µm (4.6 ‐ 13.2); OR 1.3 (0.9 ‐ 1.9); PC 14.7 µm (8.2 ‐ 16.8) 1.2 x CL (1 ‐ 1.4). Cell posterior often wide and irregular. Smooth gliding with small AC. Stationary cells make branching finger‐like and lamellar pseudopodia. Cells sometimes attached together by pseudopodia and glide or wobble. Oval scales have 4‐5 struts at each end; about 5 lateral holes per side; large end holes not circular as in Ts z. zhukovi and Page 39 of 92 40 saxoni, but a curved sausage‐shape as in Ts z. carolinensis/paracarolinensis oSL x oSW; 0.56 x 0.31 µm (0.48 ‐ 0.61 x 0.28 ‐ 0.34; N = 6); OR 1.9 (1.6 ‐ 2.1). Triangular scale (roughly equilateral) 0.48 µm (0.46 ‐ 0.52; N = 3). Type strain 4e (Woodland soil, moss and lichen, Oxford, UK. JMS. 2009). Type ITS1 & 2 rDNA sequences GenBank KF577824. ITS2 differs from nominal subspecies by an enlarged loop in helix III where CBCs are usually present. Etym: vari L. us cr ip t various squama L. scale. Comment: TEM yielded only few, often broken scales; large polynucleate cells with multiple cilia exhibited few scales. Thaumatomonas aff. seravini strain IVY18a (from an ivy leaf, Oxford, U.K. A. Howe, 2007). This strain groups with Ts seravini on trees and could be that species. However its 18S rDNA sequence (KC243106) has one significant difference (pos. 1659) from all the other 23 Thaumatomonas sequences in a conserved region, so it might a separate species, but without an scale or ITS sequence data we cannot verify that. The original T. seravini sequence AF411259 (ATCC 50636) had a missing nucleotide, probably an error as a nucleotide is present in this position in all other monadofilosan sequences including our resequenced T‐2. We briefly M describe this strain: CL 10.7 µm (7.7 ‐ 12.3; N = 14); CW 7.38 µm (4.6 ‐ 8.6); OR 1.5 (1.2 ‐ 1.7); PC 13.1 µm (11.4 ‐ 15.9) 1.2 x CL (1.1‐1.5). Smooth gliding oval to round also egg‐shaped; cell shape and size vary, sometimes quite broad with irregular asymmetric indent in posterior half. ed Stationary cells include aggregates; branching pseudopodia readily formed, varying in thickness, sometimes with eruptive broad ends. pt New Genus Cowlomonas. Diagnosis: Very fast gliding, sometimes floating, dorso‐ventrally flattened and rounded colourless heterotrophic biciliate with two cilia ce of unequal length emerging from subapical pit continuous with a ventral groove. Prominent anterior hood extending as a ridge ventrally towards the posterior on the Ac right side. Pseudopodia emerge ventrally from posterior end. Like Esquamula centrioles divergent, not parallel as in most thaumatomonads; unknown whether scales present or absent. Type species Cowlomonas planata sp. n. Type Figure 7A. Diagnosis: CL 11.8 µm (9.1 ‐ 13.6 N=4); CW 10 µm (6.8 ‐ 11.8 N=4); CL x CW ratio 1.2 (1‐1.34 N=4); AC 8.3 µm (6.8‐10 N=3); PC 20.6 µm (20.5 ‐ 20.9 N=3); PC = 1.84 x CL (1.5 ‐ 2.25 N=3). Broadly elliptical, dorso‐ventrally flattened gliding cell with distinct anterior hood, two conspicuous subapical cilia of different lengths; longer PC reflexed backwards during gliding can point forwards during pseudopodial feeding. Cell glides smoothly, very fast, with trailing PC; AC constantly beats stiffly, not undulatorily. In floating/swimming cell PC in constant sporadic motion, jerking the cell, PC repeatedly relaxes then stiffens, slightly bent at midpoint. Page 40 of 92 41 Stationary/feeding cell extends thick finger‐like, weakly branching, mainly blunt pseudopodia ventrally from sometimes visible posterior opening; cilia then inactive. Two or three highly refractile bodies up to 3 µm. Cysts not observed. Unlike Protaspa, but like Esquamula, nucleus not readily seen by light microscopy. Strain VWB1B, now dead, from marine surf, Wreck Beach, Vancouver, Canada. Etym: Cowl E. a monk’s hood, L. planus flat. Comments: Cultured using us cr ip t 50% artificial seawater, there were never enough Cowlomonas cells to sustain a stable culture and allow DNA extraction. A pico‐sized colourless chrysomonad contaminant, possibly a food source, was often attached to the gliding cells via a long stalk and trailing behind higher up in the water column; if Cowlomonas required eukaryotic food, that could explain why single‐cell picking attempts to obtain a mono‐protist culture were fruitless. It is most similar to the only known scale‐free thaumatomonad Esquamula lacrimiformis (Shiratori et al. 2012), also marine. Esquamula lacrimiformis differs in shape (teardrop rigid cell, commonly with granulated an surface) and is smaller (4.5‐11.3 X 3.9‐8.8), though size overlaps, with shorter cilia (anterior half that of Cowlomonas, posterior three‐quarters. Their behaviour is similar: smoothly gliding M cells with thick ventral pseudopodia, but the anterior cilium of Esquamula is almost motionless. We do not place Cowlomonas in Esquamula because cell shapes differ, E. lacrimiformis pseudopodia were not commonly emitted from the posterior as in Cowlomonas, and especially ed because we failed to prove the absence of scales (three attempts at TEM whole mount preparation failed even to show cells). Cowlomonas and E. lacrimiformis both have a deep ciliary pocket that continues as a ridge along the ventral side of the cell (Shiratori et al. say pt ‘groove’ although the continuation of the flagella pocket seems more a ridge than groove), but Cowlomonas is more flattened and has an obvious hood at the apical end, absent in ce Esquamula. The lack of an obvious large nucleus as in the thecofilosean Protaspa, which also combines an anterior ciliary pit and ventral pseudopod‐emitting groove, makes it less likely to Ac belong in Cryomonadida than in Thaumatomonadida. The other major flagellate group combining posterior ciliary gliding and pseudopodia is Sulcozoa (Cavalier‐Smith 2013), but we saw no evidence of a distinct dorsal pellicle, and the ciliary pit and its surrounding cowl continued on the right side by a ventral ridge is much more like Esquamula than any known sulcozoan. No sign of scales as are evident in Thaumatospina arabica (Fig. 10b) was seen by light microscopy so we place Cowlomonas in Esquamulidae; even if scales were present that position need not be wrong as sequence trees show that Esquamula lost them secondarily. Two superficially somewhat similar marine flagellates were considered as potential relatives because of their cowl‐like anterior end, ciliary pits, and posterior gliding cilium: Abollifer prolabens (Vørs 1992a) and Heterochromonas opaca (Lee and Patterson 2000). Page 41 of 92 42 However, neither has pseudopodia and Cowlomonas is thus distinct from both. Abollifer glides on its anterior cowl‐like swelling, with its rear end is raised high above the substratum. It has only one cilium (a second sometimes present is presumably predivisional). H. opaca has similar cell shape and two similarly proportioned cilia to Cowlomonas but is larger. Neither Abollifer nor Heterochromonas has been placed in a phylum; we suspect that they are mutually related, us cr ip t probably both Cercozoa, but not close to Cowlomonas; we place them incertae sedis within Ventrifilosa. New ventrifilosan orders and family Several undoubted ventrifilosan genera for which rDNA sequences are available have not been placed in orders. To rectify that deficiency we establish: an Variglissida Cavalier‐Smith ord. n. Diagnosis: scale‐free cercozoan zooflagellates with only one long posterior cilium used for gliding on surfaces; either dorsally rigid cells with ventral groove and no pseudopodia and reduced or absent anterior cilium (Clautriaviidae M Cavalier‐Smith in Cavalier‐Smith and Scoble, 2013) or soft‐bodied cells with filose branching and reticulose pseudopodia and prominent anterior cilium (Nudifilidae Cavalier‐Smith in Howe et al., 2011a). Phylogenetically most closely related to Marimonadida and Euglyphida; grouped ed with them in imbricate subclass Placonuda. Discomonadida Cavalier‐Smith ord. n. Diagnosis: as for sole included family: pt Discomonadidae Cavalier‐Smith fam. n. Diagnosis: rigid, round disc‐like dorso‐ventrally flattened gliding zooflagellates with two conspicuous cilia, with divergent centrioles sub‐ ce apically inserted behind apical notch; large nucleus conspicuous in light microscope as in Protaspa, but unlike them with no pseudopodia; glide on posterior shorter cilium. Type genus Ac Discomonas (Chantangsi and Leander 2010). Placed incertae sedis within cercozoan superclass Ventrifilosa. Though no sequences are available for Zoelucasa, its imbricate scales suggest that it is related to euglyphid filose amoebae in Imbricatea, but Nicholls (2012a) considered that it cannot be placed in an existing order. Its parallel ciliary bases, reflexed posterior cilium and large prominent nucleus are all features shared with thaumatomonads, but as its scales differ from any so far known in thaumatomonads, strongly overlap like those of euglyphids and are not known to be made in association with mitochondria, and it lacks a ventral groove and posterior ciliary gliding, we do not place it in Thaumatomonadida. The absence of pseudopodia Page 42 of 92 and presence of cilia equally strongly differentiate it from Euglyphida, so we establish a new 43 order and family for it within superorder Euglyphia: Zoelucasida Cavalier‐Smith ord. n. Diagnosis: as for sole included family: Zoelucasidae Cavalier‐Smith fam. n. Diagnosis: non‐gliding biciliates cells with reflexed posterior cilium with t parallel centrioles, large nucleus, and rigid lorica of large imbricate, probably silicified oval us cr ip scales; scales probably double‐layered, with meshwork connecting the smooth unperforated layers. Type genus Zoelucasa Nicholls, 2012. Discussion an Our results strengthen the view that scale structure and rRNA phylogeny are highly congruent in thaumatomonads (Howe et al. 2011a): the seven silica‐scaled thaumatomonad genera now sequenced group discretely and logically according to M scale morphology on our trees. Our most striking findings are of two phylogenetically distinct clades of marine former Thaumatomastix, for which we established two new ed genera, Ovaloplaca and Thaumatospina, and the discovery of Scutellomonas, which has oval plate scales like its sister Ovaloplaca, but no spine scales. Discovery of another Reckertia species confirms that Reckertia form a distinct clade. As previously pt emphasized (Howe et al. 2011a), molecular divergence in 18S rDNA between ce Thaumatomonas species is low, but by sequencing also ribosomal DNA ITS regions for numerous new strains, studying ITS2 secondary structure, and resequencing several strains we were able to sort out definitively a long‐standing historical muddle about Ac which sequences belong to which species (see Howe et al. 2011a), establish two new species and four new subspecies, put one old ‘species’ in synonymy with Thaumatomonas coloniensis, and show that Allas (with two new species) and Thaumatomonas are phylogenetically and morphologically distinct. Splitting Thaumatomastix into three genera Howe et al. (2011a) removed species without spine scales from Thaumatomastix and put them in Reckertia (Conrad 1920), of which there are now 10 species not 11, since the freshwater R. nigeriensis was shown to have eccentric spine Page 43 of 92 44 scales and moved back to Thaumatomastix (Nicholls 2012b), so Reckertia is at present exclusively marine. Our trees confirm that neither clade with spine scales groups with Reckertia, and thus fully support that narrowing of the genus. Until recently all spine‐ bearing Thaumatomastix studied ultrastructurally were marine and it had been t assumed that their spine scales were homologous with those of the freshwater type us cr ip species Tx setifera Lauterborn and that they therefore belong to the same genus. However, Nicholls (2012b) has shown that two freshwater Thaumatomastix species (triangulata, nigeriensis) have spine scales that cannot be regarded as homologous with those of marine Thaumatomastix. Beech and Moestrup (1986), who first moved C. triangulata to Thaumatomastix, gave too little weight to the marked differences in spine‐scale bases between C. triangulata and marine Thaumatomastix that we an emphasised above. Howe et al. (2011a) moved Thaumatomastix triangulata (originally Chrysosphaerella triangulata (Balonov 1980)) back to Chrysosphaerella because it M contradicted the original description of Thaumatomastix. But Nicholls (2012b) returned C. triangulata to Thaumatomastix because he found a cell with scales very similar to Balonov’s first description of C. triangulata, but which lacked chloroplasts in ed contradiction to Balonov. Though we cannot exclude the possibility that Balanov really described a Chrysosphaerella, we provisionally accept Nicholl’s (2012b) assumption pt that Balanov made several mistakes when describing it and interpretation of this species as a freshwater Thaumatomastix with scales very similar to Tx nigeriensis. ce The type species Thaumatomastix setifera of Lauterborn (1896, 1899) had spine scales and a prominent 1 μm thick layer at the surface of the cell resembling Ac small boxes (Lauterborn 1896, 1899), which we interpret as the plate‐scale layer in side view. The LM images and drawings of Tx nigeriensis (Nicholls 2012b) resemble the type description of Thaumatomastix setifera, for which ultrastructural detail of the scales can never certainly be known. All three freshwater species, Tx nigeriensis, Tx triangulata, and setifera, exhibit a distinct eccentric nucleus, spine scales, boxed cell boundary layer, and pseudopodia. All other former Thaumatomastix species are marine, and all except Thaumatospina tripus have plate scales with the two tiers separated by ~0.5 μm, not 1 μm as in freshwater species. Page 44 of 92 45 Howe et al.’s (2011a) argument that C. triangulata should not be in the same genus as marine Thaumatomastix with spines remains completely sound. Balonov’s pictures of Tx triangulata show eccentric spines coming from a bobbin base nearer the size of the plate scales, which is entirely different from the marine ‘Thaumatomastix’ t species (now Thaumatospina and Ovaloplaca), which have radially symmetric spine us cr ip bases. Our restriction of Thaumatomastix to the three freshwater species and establishing those new genera now completely reconciles that argument of Howe et al. (2011a) with Nicholl’s (2012b) reinterpretation of Tx triangulata. Before Nicholls (2012b), only the plate scales of Tx triangulata had been reported (never spine scales) since its first description (Thomsen et al. 1993; Vørs 1992a) so our present reinterpretation of spine‐scale base evolution was impossible before Nicholls an discovered its spine scales and showed they were morphologically like those of Tx nigeriensis, and very unlike those of any marine ‘Thaumatomastix’. We therefore M disagree strongly with the assumption that Thaumatomastix sensu stricto spine bases are homologous with those of marine species, as implied schematically in Thomsen et ed al. (1993). We explain below how we think these two grossly dissimilar spine‐scale morphologies evolved independently from plate‐scale ancestors by modifying their pt biogenesis in radically contrasting ways. Though no sequences are known for any freshwater Thaumatomastix, we predict that they will not branch with either marine ce clade, but somewhere within the more derived freshwater clade comprising Thaumatomonas, Allas, and two deep‐branching environmental DNA lineages, either Ac (or both) of which could turn out to be genuine freshwater Thaumatomastix. In both freshwater species plate‐scales are triangular with rounded corners, somewhat like those of Thaumatospina or Reckertia (Nicholls 2012b). Scale evolution within Thaumatomonadida Each major thaumatomonad clade has a distinct broadly uniform scale morphology, which maps logically onto the trees (Fig. 2). The main limitation for reconstructing the history of their scale morphology is that we do not yet know where Gyromitus fits on the tree. This is important because its scales are oval latticed hollow Page 45 of 92 46 cylinders with strongly outrolled upper and lower rims and a sieve‐like sheet across its lower tier, markedly different from both the two‐tier plate scales and the flanged spine scales of Thaumatomonadidae. As all three of them and the morphologically distinct but somewhat similar scales of Peregrinia, are finally made in mitochondria‐associated us cr ip we suggest that all four diverged from a common ancestor in ancestral t vesicles (Mylnikov and Mylnikov 2012; Ota et al. 2012; Swale and Belcher 1974, 1975), thaumatomonads that first evolved this mitochondrial association, which is found only in thaumatomonads. Gyromitus is currently classified in Peregriniidae (Howe et al. 2011a) and it is possible that environmental sequence Amb_18S_1066 that is sister to Peregrinia is a Gyromitus rather than another Peregrinia. The structure and symmetry of the upper and lower tier connections are so substantially different (Fig. 11) that we an do not accept the inclusion of Peregrinia within Gyromitus by Karpov (2011). All trees indicate that Esquamula is a peregrinid that has lost scales. Our PhyloBayes tree (Fig. M 1.1) suggests that Spongomonadida may also be derived from thaumatomonads by scale loss and is consistent with the idea that silica scales evolved in the common ancestor of Thaumatomonadida and Euglyphida – the ancestral imbricate (Cavalier‐ ed Smith and Chao 2003) and were lost one more time in the common ancestor of Nudisarca, which are apparently sisters of Euglyphida. pt The simplest interpretation of silica scale evolution in imbricates is that scales were ancestrally single‐tier plate‐scales made in the Golgi complex, as in Euglyphida. ce Fig. 11 presents a unifying explanatory hypothesis for scale evolution and development in imbricates and how the major kinds of imbricate scales may have diverged from a Ac common ancestor. We suggest that the lower tier of plate‐scales in Thaumatomonadidae and the proximal sheet in Peregriniidae are homologous to each other and to the single‐tier euglyphid scale. We postulate that even today in Thaumatomonadida lower‐tier scale rudiments first assemble in Golgi vesicles and that these are then moved to the mitochondria for completion and adding the upper tier (Ota et al. 2012; Zolotarev et al., 2011). We suggest that association with mitochondria is primarily a shaping device for the upper tier only, and might have been unnecessary for the ancestral first steps of lower tier assembly. Consistently with this hypothesis, developing two‐tier plate scales are always oriented with the upper tier facing and Page 46 of 92 47 projecting into the invaginated mitochondrial envelope (Mylnikov and Mylnikova 2012; Ota et al. 2012; Swale and Belcher 1974, 1975). Gyromitus plate scales do not have a definite upper tier and the upper flange is connected to the lower tier by a continuous lattice, not by discrete widely separated us cr ip t struts or groups of struts as in all other genera; both features are plausibly the ancestral condition for thaumatomonad plate‐scales as they could simply have evolved by extending the ring into the shaping mitochondrion. In Peregrinia, the upper tier lacks a central sheet and consists of laterally inrolled sides supported by terminal struts. The spine‐scales with central spine and narrow flange found in Thaumatospina and Ovaloplaca only clearly evolved in the last common ancestor of an Thaumatomonadidae after it diverged from the Peregriniidae/Esquamulidae clade that lacks spine‐scales, and were lost in the common ancestor of the Reckertia/Allas/Thaumatomonas clade and independently by Scutellomonas. We M consider the flange as homologous with the rim of the continuous upper tier in body plate‐scales and that it became narrower than the lower tier when the central region ed of the organic precursor material was pushed out to form the hollow spine. The fact that the flange is connected to the base plate by 3‐fold symmetric struts in these radially symmetric spines strongly suggests that plate‐scales at the time of this pt differentiation were themselves triangular with three symmetrically placed corner struts which evolved directly into the spine‐scale struts. In Ovaloplaca salina (Ota et al. ce 2012) the three struts are all clearly basally dichotomous, as they also appear to be in Thaumatospina vancouveri (Fig. 6B), mexica (Fig. 6C), and arabica (Fig. 7C); the base of Ac each spine‐scale strut is also clearly flared outwards in Ovaloplaca yabukii (Fig. 8F), but moderately so in O. asymmetrica and scarcely so in O. multipora (Nicholls 2012b). The single corner struts of all three Thaumatospina described here can also be interpreted as basally dichotomous. We therefore suggest that each subflange strut in Thaumatospina/Ovaloplaca is homologous with and evolved from a single basally dichotomous corner strut of triangular spine scales present in the common ancestor of Thaumatomonadidae. In Reckertia gemma the single corner struts of the triangular scales are also basally dichotomous (Fig. 8H), as they also are in R. filosa (Howe et al. 2011 fig. 7B4), Page 47 of 92 48 though interpretation is complicated by the presence of a third density that may not be basal which makes them look tripartite in surface view; their fig. 4B5). The micrographs for R. dybsoeana have insufficient resolution to show whether the strut base is merely flared or truly dichotomous (Thomsen et al. 1993); those of R. formosa t are somewhat different with an upper part resembling an inverted hollow cone that us cr ip underlies the upper tier corner perforation and a narrow proximal portion that is only slightly flared at its base. We suggest that Reckertia retained triangularity of its plate‐ scales from the ancestor of Thaumatomonadidae with the same type of basally dichotomous strut, and that its triangular scale architecture is homologous with that of Thaumatospina plate‐scales. On that interpretation the oval plate‐scales of the Ovaloplaca/Scutellomonas clade (mutually similar in both genera, but markedly an different from the oval scales of Thaumatomonas and Peregrinia) evolved secondarily from such triangular scales by addition of a fourth basally dichotomous/flared strut: in M Scutellomonas the structure is clearest – one basally flared strut at each end of the oval and one at each side at its pinched‐in waist. ed Reckertia species have triangular scales (curved vertices) of two seemingly distinct morphologies. Most similar to the triangular plate scales of Thaumatospina (Ta triangulata, Ta tripus, Ta patelliformis) are scales with a partially sheet‐like upper tier pt joined to the bottom tier by large hollow 'wine glass' struts (R. formosa) or by short rods at the corners (R. filosa). These differ from Thaumatospina plate scales primarily ce by having three prominent corner holes located just centrally to the corner struts (R. filosa) or over their wine‐glass cavity (formosa); some species also have less Ac conspicuous elongated lateral holes (elongated in Ta filosa; possibly not present in Ta formosa). Superficially quite different are scales with a very open, latticed upper‐tier, often with a dense crown‐like central structure raised on longer varyingly slender supports that meet the sides of the triangle, not the vertices as do inter‐tier struts (e.g. R. gemma, R. sagittifera, R. spinosa, R. dybsoeana, R. splendida). As the crown‐scale R. gemma is closely related to R. filosa with sheet‐like upper tier (Fig. 1.1) there is no need to place them in separate genera, and the differences must be able to evolve relatively simply. Page 48 of 92 The R. filosa type with sheet‐like upper tier is more like that of outgroups, 49 suggesting that the open crown type may have evolved from it simply by greatly expanding pre‐existing holes in the upper tier, leaving only slender supports after such expansion. Enlarging corner holes only by a more moderate amount than in most t species would produce an intermediate structure with a central triangular sheet very us cr ip like that of Fig. 3A of Gooday et al. (2006), misidentified as Tx dybsoeana; as well as its central sheet being broader than in R. dybsoeana, Fig. 3A shows two additional curved loops interconnecting the vertices across the corner holes unlike in R. dybsoeana or any Thaumatospina: though distinct from all described Reckertia its multiple perforations suggest membership of that genus ‐ as this was a single isolated scale there is no evidence that it donor organism had separate spine‐scales like an Thaumatospina. Still further enlargement of corner holes only would yield an even more open structure like that of Gooday et al. Fig. 3B with a broad residual central 3‐ M armed structure, or with slightly less than that ones like Fig. 14 of Thomsen et al. (1993). Enlarging lateral holes only and pushing the central region out distally would yield structures like species 1 and 2 of Thomsen and Ikävalko (1997), whereas ed substantially enlarging both corner and lateral holes might yield R. fusiformis and fragilis (though interpreting their often broken structure is hard). Enlarging both types pt of holes would yield a suitable precursor of R. gemma scales but a third tier would need to be added to make the jewel. Interpreting Fig. 3A,B of Gooday et al. 1993) is ce difficult from a single scale seen in aerial view, as one or both might have lower tier perforations also contributing to the observed structure, unlike all described Ac thaumatomonads. R. splendida from Antarctic and Arctic sea‐ice (Thomsen et al. 1995; Thomsen and Ikävalko 1997) was removed from Thaumatomastix (Howe et al. 2011a) because unlike Thaumatospina it has only one scale type with a large two‐tiered triangular base and central ‘spine’ with three very long dichotomously‐based struts but no flange and which connect to the side of the rail‐like marginal band of the upper tier of a plate scale; it is thus not structurally homologous with spine‐scales of Thaumatospina/Ovaloplaca. Its superficially spine‐like central projection, extends from the junction point of the three lateral supports, which are probably homologous with Page 49 of 92 the crown‐like tripartite central structures of ‘Tx’ species 1 and 2 in Thomsen and 50 Ikävalko (1997), both probably undescribed Reckertia; the R. splendida central projection is terminally bulbous and internally tripartite, and probably homologous with the terminal point of such tripartite crowns of other Reckertia species; it almost t certainly elongated entirely independently of the differently tipped spines of us cr ip Thaumatospina/Ovaloplaca and Thaumatomastix; unlike them it is not a second differentiated scale type but a simple modification of a multiperforated plate scale. The triangular hole that makes the base of each lateral support dichotomous in R. splendida is probably homologous with the lateral holes of R. filosa; the three lateral supports could be the residues left after a vast expansion of corner holes. R. gemma arguably duplicated this dichotomous lateral structure when evolving the jewel layer, an so unlike most latticed Reckertia it has 6 forked lateral supports not just three. The unusual scale diversity of Reckertia is consistent with its being a relatively M deep‐branching clade (compared with Thaumatomonas/Allas) likely to be over a hundred million years old. We suggest that the entire Reckertia‐containing clade is ed probably Reckertia; it will be interesting to determine how its varied scale types map across this clade. R. splendida and most other Arctic species (Thomsen and Ikävalko 1997) might belong in the sea‐ice environmental subclade R1 (Fig. 1.1). Figure 11 pt assumes that upper‐tier corner holes evolved just once in the common ancestor of Reckertia and Allas/Thaumatomonas and were secondarily lost in Reckertia igloolika, ce which can be tested by establishing its position on the sequence tree. It is noteworthy that both sequenced non‐polar Reckertia (filosa and gemma), which are closely related Ac on Fig. 1.1, have oval ciliary scales; conceivably therefore Reckertia with round ciliary scales (e.g. R. formosa, R. dybsoeana: Thomsen et al. 1993) are more distant and might branch in sister clade R1. However, though the micrographs are less clear, it appears that R. dybsoeana’s jewel may constitute a third tier as in R. gemma, in which case it is perhaps more likely to be phylogenetically closer to it than to the other Arctic species; possibly these other two three‐tier open lattice scale types evolved from a sheet‐like ancestor independently of the others that lack a third‐tier domed gem. The presence of oval ciliary scales in both sequenced Reckertia and in Ovaloplaca suggests that oval Page 50 of 92 51 ciliary scales may have been ancestral for the whole clade, which might have become secondarily circular and smaller in Arctic species. In contrast to Thaumatospina, when Thaumatomastix eccentric spines evolved later, the spine grew not from the central area but from the rim to one side of a us cr ip t central hole (which if present when it originated would have prevented central spine formation); in this spine type the plate‐scale upper tier remained the same diameter as the lower tier and was not narrowed into a flange. These key differences strongly suggest that the two types of spines evolved independently from two‐tier plate‐scales with originally equal‐sized upper and lower tiers by separate morphogenetic processes. It is hard to see how the eccentric Thaumatomastix spine could have evolved directly an from the Thaumatospina/Ovaloplaca type. The much smaller single‐tier oval scales on the anterior cilium of Ovaloplaca M and Reckertia only could have evolved in their common ancestor after it diverged from Thaumatospina by interrupting the developmental pathway of body scales while the rudiment was still very small; this might in some cases happen before they left the ed Golgi for adding the upper tier in the mitochondria, in which case these ciliary scales might pass directly from the Golgi to the ciliary base without a mitochondria‐ pt associated phase. However, in one Reckertia even ciliary scales may associate with mitochondria (Moestrup 1982). Though ciliary scales show much less variation than ce body scales, those of ‘Thaumatomastix’ sp. strain TM‐5 (Zolotarev et al. 2011), which is the very same strain described as Reckertia filosa by Howe et al. (2011), has a Ac prominent distal comb not seen in some others, which might well require mitochondrial shaping. Ciliary scales were probably lost independently by Scutellomonas that drastically shortened its anterior cilium and by the ancestor of Thaumatomonas/Allas/Thaumatomastix that retained a relatively long one. In the common ancestor of the Allas/Thaumatomonas clade lateral and morphologically distinct corner or end perforations evolved in the strongly upwardly bowed upper tier which is at least somewhat smaller than the lower tier. As its outgroup, Reckertia, has triangular scales it is likely that the Allas/Thaumatomonas ancestor also had triangular scales. Allas itself uniformly has scales with an oval base Page 51 of 92 but rounded triangular upper tier supported by three dichotomously based struts, 52 which could have evolved simply by squashing the upper tier laterally during development. The ancestor of Thaumatomonas however arguably evolved oval scales differently by suppressing altogether one of the three ancestral struts and keeping one t only at each end. The Ts seravini/zhukovi clade’s ancestor clearly kept both such two‐ us cr ip strut oval scales and three‐strut triangular scales, but some descendant lineages lost the triangular version and others lost the oval ones. Within Ts zhukovi, some strains retained the putatively ancestral single broad terminal struts as in triangular scales, but some subdivided these into a larger number of narrower struts (as detailed in Results). Interestingly, scale development in the Ts zhukovi type strain T‐3 still seems developmentally or evolutionarily labile in culture, as when it was first isolated it had an both triangular and oval scales (and irregular ones, some even with four vertices), but a decade later we saw only triangular ones; too few cells were examined to determine M whether that difference reflects developmental variation or a rapid evolutionary change. An analogous example of rapid ultrastructural changes (eyespot degeneration) (Moldrup et al. 2013). ed during prolonged laboratory cultivation was recently noted in a dinoflagellate By contrast the ancestor of the Ts lauterborni/solis clade evolved multiple pt terminal struts and at the same time lost the triangular version of the scale, which seems to have re‐evolved in just one lineage; one explanation for that is that the ce pattern formation process for Thaumatomonas plate scales can in principle produce either three struts or two struts depending on tweaking some quantitative parameter Ac (e.g. diffusion rates or binding interactions of morphogens or inhibitors), which can be relatively easily changed or reversed in evolution. The evolution of multiple narrow end struts instead of just one broad one might also be a consequence of increasing its dimension and thus easily evolve in parallel. The broad single struts of Thaumatomonas and Allas appear as a less dense ‘curtain’ connecting their more electron‐dense curved edges. The curtain is clearly seen in A. multipora and A. media (Fig. 9B and D, respectively) and Ts zhukovi strain UPL1Be2 (Sup. Fig. 6G). Multiple end struts are clearly separated structures that appear as black dots from above, as seen in Page 52 of 92 53 Ts zhukovi strain GMKL4 (Fig. 10D), Ts oxoniensis strain EP3 (Sup. Fig. 3E) and Ts solis (Fig. 5D). Our interpretation of Thaumatomonas scale evolution makes it likely that the PhyloBayes tree showing the Ts lauterborni/solis clade, as holophyletic is correct. It is us cr ip t simpler to suppose that end struts multiplied only once in the ancestor of this clade than separately in Ts solis as would be the case if the trees by other methods that place Ts solis more deeply before the divergence from the Ts zhukovi/seravini clade were right. The correct identification here of all sequenced Thaumatomonas and much improved sampling has enabled a much simpler and more comprehensive interpretation of Thaumatomonas scale evolution than was previously possible. an Isolation of two new Allas species with the same general kind of scale as A. diplophysa, with non‐equilateral triangular upper tier and oval lower tier, both distinct in upper tier perforation detail, confirms our earlier view that Allas and Thaumatomonas can be M reasonably treated as distinct genera (Howe et al. 2011a); our PhyloBayes trees, which uses the most evolutionarily realistic heterogeneous model, suggest that they are ed sisters and each is holophyletic. Thaumatomonas coloniensis MLTB12 from a chalk cliff in Malta is the only pt strain in the lauterborni/solis clade that apparently has triangular scales and oval scales with broad struts, more like those seen in the zhukovi clade, as well as what ce could be the usual oval scales (Sup. Fig. 4D). We do not revise the description of Ts coloniensis to include the greater variety of scales found in our single TEM preparation Ac of strain MLTB12 because it was technically poor and no repeat preparation could be made to verify the scale types. As we did not isolate the culture ourselves we also cannot be sure it was clonal or say whether scale morphology changed since isolation. The numerous Thaumatomonas strains studied show that cells with the same 18S rDNA can produce different scale types, indicating that in this genus scale morphology can evolve faster than 18S rDNA genes and be extremely discriminating taxonomically. Morphological differences between thaumatomonad genera All thaumatomonad genera, as now defined, group distinctly on trees (Fig. 1.1 & Fig. 2), and show clearly correlated morphological differences. Reckertia and Page 53 of 92 Ovaloplaca are the only thaumatomonads with single plate scales on their anterior 54 cilium. (The minute dense square‐ish knob‐like structures called ‘cone‐shaped scales’ by Mylnikov and Mylnikova (2012) seen on both ciliary surfaces in the type strain of Ts zhukovi zhukovi are not plate‐scales, and are apparently also present on the plasma t membrane lining the ciliary pocket; they are components of the cell surface coat that us cr ip might better be called minute knobs than scales; they are clearly not homologous with true scales of other thaumatomonads. Reckertia cells are oval to round and more dorso‐ventrally flattened than Thaumatomonas cells, which appear more convex. Thaumatospina cells are inflated, not flattened, and a long AC (up to half cell length) and spine scales are visible in the light microscope, and readily produce pseudopodia. Scutellomonas has a very distinct bulbous anterior end, no visible AC and is the fastest an glider, along with Cowlomonas, which is distinctive in producing ventral pseudopodia from a posterior opening. In Cowlomonas the long AC flicks to one side of the cell. The M AC for Thaumatomastix Lauterborn (1896, 1899) was almost the same length as the posterior. The AC in Ovaloplaca yabukii and O. salina (Ota et al. 2012) is very short (barely visible), further support for their being a different genus from Thaumatomastix ed and Thaumatospina. The AC of Thaumatomonas and Allas are always a lot shorter than Thaumatospina and some published Reckertia; however the Reckertia in this study pt also have a short AC and glide faster than Thaumatomonas, and have an obvious flagella pocket (notch) off‐centre. Therefore AC length varies considerably in Reckertia. ce Despite the above differences, not all thaumatomonad genera can be distinguished unambiguously in the light microscope, the Allas/Thaumatomonas Ac distinction especially, though all genera and most species can be recognised by electron microscopy. There is inconclusive evidence that Allas may not produce pseudopodia as readily as Thaumatomonas, but they cannot safely be distinguished in the light microscope. Within Thaumatomonas DNA sequences also are necessary for clearly distinguishing the closely related species in which scale variation is high. All three species of Allas now described have uniform two‐tier scales with oval lower tier and a non‐equilateral triangular upper tier supported by three broad struts that flare dichotomously at their base: one at each end and one on the side. Those of Thaumatomonas are much more variable, the majority being oval with oval upper tier Page 54 of 92 and only end struts, but in some strains both tiers are triangular with a single broad 55 basally dichotomous strut at each vertex. Most strains in the Ts coloniensis/lauterborni/solis clade have only oval scales, but one Ts coloniensis strain (ML2B12) has non‐equilateral triangular scales as well. Ts seravini always has t equilateral triangular scales, either alone (type) or mixed with oval two‐strut scales. us cr ip Thaumatomonas zhukovi always has 2‐strut oval scales and seven of the nine Ts zhukovi also have at least a few non‐equilateral triangular or irregular scales. None of these Thaumatomonas scale types is the same as those of Allas, which unlike any Thaumatomonas scales always combine oval lower tier and triangular upper tier. Some former Thaumatomastix spines are hollow and some are winged, but this an is only really visible when using TEM not SEM, which can hide structural internal details. The triangular body‐scaled Thaumatomastix now called Thaumatospina (Ta tauryanini, tripus, patelliformis, vancouveri, arabica, mexica) all have hollow‐shaft M spine‐scales (Mikrjukov 2002; Takahashi and Hara 1984). However, oval scale former Thaumatomastix, now called Ovaloplaca, may have either winged spine‐scales with a ed hollow proximal half (O. salina) or smooth hollow shaft spine‐scales with bifurcate distal ends, O. biparta (Beech and Moestrup 1986; Thomsen et al. 1993). pt A difficulty in distinguishing Reckertia scales from the triangular variety of Thaumatomonas and Thaumatospina was exemplified by Nicholls (2012b) who noted ce resemblance between Ts hindoni and Thaumatomastix sp. #5 (figure 40 ‐ 43 of Thomsen and Ikävalko (1997), which we considered an unidentified Reckertia (Howe et Ac al. 2011a). Thaumatomastix sp. #5 has two plates of nearly equal size connected at the corners by hollow, short tubes; one plate has submarginal densely spaced perforations. Ts hindoni had a base plate nearly the same size as the upper tier with no hollow tube struts, just joins at the corners. Ts hindoni has a bowed upper tier with straight edges (surface view) similar to our Ts seravini (T‐2) strain: Figure 4F, comparable with that seen in Nicholls’ (2012b) Figure 10C; far right scale side view bows upward making visible the corner strut. Nicholls (2012b) did not report cilium scales. All Reckertia have anterior cilium scales whereas Thaumatomonas do not. Thus all characters support the transfer of R. hindoni to Thaumatomonas. Thaumatospina also might potentially be misidentified as Reckertia in the absence of spine scales or cilia, because we found that Page 55 of 92 Ta mexica sp. n. sometimes produces cells with only triangular scales despite other 56 cells from the same clonal culture having both triangular and spine scales (Fig. 8D and 8F). However, Reckertia all have anterior cilium scales, unlike Thaumatospina, a clear distinction between then genera. us cr ip t Thaumatomonad posterior cilium acronematic tip Four thaumatomonad genera, Scutellomonas, Thaumatospina, Allas and Thaumatomonas, displayed variable length acronematic posterior cilia. Acronemes have apparently not been described before for any thaumatomonad. Scutellomonas has a prominent thin acroneme that sticks to the substratum causing it to stretch before retracting back with the gliding cell. Thaumatospina (Fig. 7A), Allas and an Thaumatomonas (Fig. 5A) do not have as long or prominent and thin acroneme as Scutellomonas (Fig. 8A); it instead appears more flattened and flared or bifurcate. In M other protists permanent acronemes often arise when ciliary central pair microtubules extend further than the outer doublets, differences in the length of such protrusion accounting for different acroneme lengths. Temporary acronemes can arise in gliding ed flagellates if the membrane tip sticks to the substratum when the rest of the cilium is moved forwards by kinesin motors interacting with the microtubules during gliding, pt which would leave an elastic membranous tip empty of microtubules whose length would depend simply on the lag between membrane de‐adhesion and forward gliding. ce Variable lags in de‐adhesion could explain the variable acroneme length and sudden retraction seen here and in some other gliding flagellates, notably the sulcozoan Ac Mantamonas plastica (Glücksman et al. 2011). Improving monadofilosan phylogeny Of key importance for better understanding imbricate scale evolution is to determine by sequencing whether Zoelucasida and Rotosphaerida really belong in Imbricatea. It is also important to study their scale development to see if they are made in association with mitochondria as in Thaumatomonadida, or not as in Euglyphida. Through our improved understanding of thaumatomonad plate‐scale homologies and evolution summarised in Figure 11, we have become sceptical of the idea that the diverse scales of rotosphaerids are more closely related to Page 56 of 92 57 thaumatomonad two‐tier strutted scales than to those of other imbricates (Cavalier‐ Smith and Chao 2012). Whilst it remains reasonable to consider rotosphaerids as imbricates, a direct relationship between Rotosphaerida and Thaumatomonadida seems more questionable. Because of its parallel ciliary bases we place Zoelucasida in t subclass Placoperla (within superorder Placofila because of its scales rather than us cr ip perles) not with euglyphids in subclass Placonuda; none of the flagellate members of the superorder Nudisarca (Cavalier‐Smith and Chao 2012), the apparent sister group to Euglyphida on the PB tree (Fig. 1.1), has the combination or two parallel ciliary bases with one cilium directed forwards and one backwards shared by Thaumatomonadida and Zoelucasa. Ciliary and nuclear features of Zoelucasa suggest an affinity with thaumatomonads, and it is possible that their apparently double‐layered scales are an related to those of the peregrinid Gyromitus despite the notable absence of the second layer in the central region and much thicker rim in Gyromitus. Though such a M relationship is speculative, the already known phenotypic diversity in the putative Peregrinia/Esquamula/Spongomonas subclade is sufficiently great to leave open the possibility that Zoelucasa might belong in or near to unidentified marine flagellate ed subclade MF1 (Fig. 1.1). Though another plausible candidate for a clade in that position is the aciliate filose Perlofilida (Acanthoperla, Pompholyxophrys), which like pt spongomonads (putative sister of clade MF1 on the PB tree) secrete a surface layer of silicified perles, and also require sequencing to verify their inclusion in Imbricatea and ce Placoperla, one member of MF1 (5B) is known to be a flagellate, which tends to argue against this (though not decisively as cercozoan clades can have mixtures of flagellates Ac and filose amoebae). No previously published large‐scale monadofilosan tree used the evolutionarily more realistic heterogeneous CAT‐GTR gamma models, as in our PhyloBayes CAT‐GTR 18S rDNA tree, which is also by far the most comprehensive taxonomically to date for Monadofilosa other than metromonads. Our trees also gave apparently improved phylogenies in several respects, notably support for the holophyly of the imbricate subclasses Placonuda and Placoperla, superorder Nudisarca, and new order Variglissida as distinct holophyletic groups, as well as the previously controversial distinctness of Neocercomonas from Cercomonas and the affinities of several Page 57 of 92 environmental DNA clades. For the first time our CAT‐GTR tree suggests that 58 spongomonads may have evolved from thaumatomonads by scale loss, and that pansomonads may have evolved from glissomonads. In view of the established ability of CAT‐GTR models to reduce long‐branch artefacts compared with homogeneous t models (Lartillot et al. 2007, 2008; Philippe et al. 2009; Roure et al. 2013), these novel us cr ip findings, which are important for cercozoan phylogeny though only weakly supported on single‐gene trees, need testing by multigene analyses. The positions of Discomonadida and Ventricleftida within Ventrifilosa, and of Sainouroidea all remain unclear, so multi‐gene trees will be required to determine whether each is closer to Imbricatea or Thecofilosea. an Clades X and T ‘glissomonads’ which together form one of the two robust subclades (the other being clade Y) which are sister to pansomonads on our PhyloBayes tree have now been identified as a new family of algivorous M amoeboflagellates, Viridiraptoridae (Hess and Melkonian 2013). Their amoeboid phenotype and position on our PhyloBayes tree as sister to pansomonads rather than ed to any established (less amoeboid) glissomonad group suggests that rather than assigning them to Glissomonadida (Hess and Melkonian 2013) they might more appropriately be considered early diverging pansomonads, if this alternative phylogeny pt proves to be correct. Alternatively, because viridiraptorids resemble some of the disparate flagellates assigned to Pseudospora and have been suggested to be fairly ce closely related to certain biciliate species of that possibly polyphyletic genus (Hess and Melkonian 2013), they could equally well – even preferably – be assigned to the still Ac older alga‐eating order Pseudosporida (Cavalier‐Smith 1993), whose diagnosis they fit much better than the previous diagnosis of either Glissomonadida or Pansomonada. Accordingly we here assign clades T/X and Y all to Pseudosporida, and transfer that order from Endomyxa (where it previously was in the absence of sequence evidence for it being either an endomyxan or filosan merely because of its life‐style resemblance to vampyrellids (Bass et al. 2009), which are Endomyxa (Hess et al. 2012); earlier it was in Filosa: Cavalier‐Smith and Chao 2003). If the Fig. 1.2 phylogeny is confirmed by stronger evidence, it would probably be desirable to reduce Pansomonadida and Pseudosporida both to suborders within a broadened Glissomonadida. The putative Page 58 of 92 pansomonad/pseudosporid subclade seems to have become secondarily more 59 amoeboid than in ancestral glissomonads; however until clade Te is also identified phenotypically we cannot be sure that the deepest branching glissomonads were not also more amoeboid in their trophic phase. When the e‐sarcomonad clade is t characterised it may prove appropriate to expand Glissomonadida further to include it us cr ip also. The discovery of this green‐alga eating sarcomonad clade emphasises that our knowledge of cultured sarcomonads is currently strongly biased towards bacterivorous species. Some other novel clades might also require a eukaryote food supply before they can be cultured. The culture techniques of this paper were similarly biased an towards bacterivorous thaumatomonads; future research should bear in mind that some uncultured thaumatomonad lineages may also feed on algae or other eukaryotes, as postulated above for Cowlomonas and Thaumatomastix tripus as least appears to M eat diatoms (Thomsen et al. 1993), so more complex culture procedures deliberately supplying potential algal, fungal, or protozoan foods may be necessary to characterise ed them. Brabender et al. (2012) claimed that bootstrap values are too low to support pt the monophyly of Neocercomonas (Cavalier‐Smith and Karpov 2012). However, its distinction from Cercomonas with our well‐aligned 1785‐nucleotide positions is ce unambiguous. Brabender et al. (2012) found Cercomonas and Neocercomonas intermixed on their tree, presumably an artefact arising because their alignment of Ac only 1069 nucleotides excluded too many informative positions; it was unwise to use so few for a cercomonad tree as far more can always be accurately aligned with sufficient effort. Their paracercomonad tree used 1539 positions; even so, our alignment seems better as our tree had only two contradictions between Bayesian and ML trees (the two most basal branches where our ML tree differed in topology from theirs). Their tree had at least six such contradictions; probably many more as they omitted such details for seven insignificantly supported ‘clades’, most not even present on our much better supported trees. Unlike theirs, none of our trees put Brevimastigomas in a clade sister with Metabolomonas. Their Bayesian tree did not even recover monophyly of Paracercomonas or two of its subclades (B1a, B1b); all Page 59 of 92 60 three were strongly supported (maximally for B1b) on all our trees. Even within these subclades the branching order on our more convincingly supported trees was different. Habitat conservatism and monadofilosan environmental sequences t Our trees show that thaumatomonads very rarely shifted habitat between us cr ip marine and fresh water. Reckertia, Thaumatospina, Ovaloplaca, and Scutellomonas have been found only in marine environments. The Thaumatomonas/Allas clade is wholly from fresh water or soil and groups robustly with two deeper branching environmental clades (Fig. 1.1), either of which could be (or be related to) the missing freshwater Thaumatomastix species. The rest of the thaumatomonad tree is marine except for Peregrinia clavideferens and its freshwater environmental sister sequence. an This small freshwater clade is sister to the scaleless marine thaumatomonad, Esquamula lacrimiformis (Shiratori et al. 2012). The marine environmental clade sister M to Reckertia, ‘R1 sea ice clade’ (Fig. 1.1), is a prime candidate to hold species like R. splendida from sea ice (Thomsen et al. 1995). It seems that thaumatomonads were ancestrally marine and freshwater/soil was colonised only twice ‐ if Spongomonadida freshwater. ed really nests phylogenetically within Thaumatomonadida they are a third transition to pt We suggest that imbricate subclass Placonuda was also ancestrally marine; its superorder Nudisarca is entirely marine whereas Euglyphida can be marine or (usually) ce freshwater. Combining the extensive euglyphid tree of Heger et al. (2010) with our more complete outgroup tree suggests that one major exclusively freshwater Ac euglyphid clade of three families (Euglyphidae, Assulinidae, Trinematidae) colonised freshwater early on before it diversified. The other two clades, Paulinellidae and Cyphoderiidae, remained marine initially; within each the ancestor of perhaps only one subclade much later changed to freshwater habitats (Paulinella and a Cyphoderia subclade). Thus, on present evidence, Imbricatea were ancestrally marine but colonised freshwater at least six times. Their immediate outgroups, Discomonadida and Ventricleftida, are exclusively marine, strengthening the idea of a marine ancestry. A nudisarcan position for marine sequence 9‐2.2 close to marimonads, as observed here, is more reasonable than the previous (unsupported) position within the virtually Page 60 of 92 61 entirely soil/freshwater glissomonads (Howe et al. 2011a), which was probably a long‐ branch artefact. The more distant Thecofilosea were probably also ancestrally marine, as most still are; Rhogostoma colonised soil and Pseudodifflugia is said to be marine, us cr ip t freshwater or soil. Thus, Ventrifilosea were almost certainly ancestrally marine. By contrast, virtually all Sarcomonadea are from soil or fresh water. We suggest that Monadofilosa were originally marine and that there may have been no more than about 10 transitions to soil or fresh water, and scarcely any switches back to a marine habitat. Our present estimate of seven marine to freshwater transitions if Sarcomonadea are holophyletic, and eight if they are paraphyletic as many 18S rDNA an trees marine trees weakly suggest (Howe et al. 2011a), is probably somewhat too low as sampling is incomplete. Orders Perlofilida and Rotosphaerida (both probably imbricates: Cavalier‐Smith and Chao 2012) have not yet been placed on the 18S rDNA M tree. Within Sarcomonadea only one cercomonad has been isolated from a marine habitat (Paracercomonas marina; Karpov et al. 2006) but even that grows better in ed fresh water; it is unclear whether detection of some cercomonad DNA in littoral marine samples simply reflects the washing of cysts from land to sea or whether they can actually grow in seawater – most strains tested cannot, though tolerance of low pt levels of salt is widespread (Bass et al. 2007). Distribution on our tree of the numerous environmental sequences is consistent with marine to freshwater transitions being ce distinctly rare in monadofilosan evolution. Similar saline versus non‐saline habitat conservatism over hundreds of millions of years is general in other protists (see Bass et Ac al. (2009), Cavalier‐Smith and Chao (2012) and Heger et al. (2010)). Acknowledgements We thank Denis Tikhonenkov and Alexander Mylnikov for strains and an unpublished sequence; Keith Vickerman for strains; Flemming Ekelund for information; Pavel Škaloud for help with ITS2 analysis; Takashi Shiratori and Ken Ishida for pre‐ publication data for Esquamula and Scutellomonas strains; and Akinori Yabuki for Page 61 of 92 permission to include his unpublished sequence and micrographs of Ovaloplaca 62 Ac ce pt ed M an us cr ip t yabukii; BBSRC for a studentship (JMS); and NERC for a research grant. Page 62 of 92 63 References Balonov, I.M., 1980. On the new species of the genus Chrysosphaerella (Chrysophyta). Bot. Zh. 65, 1190‐1191. Bass, D., Cavalier‐Smith, T., 2004. Phylum‐specific environmental DNA analysis reveals us cr ip t remarkably high global biodiversity of Cercozoa (Protozoa). Int. J. Syst. Evol. Microbiol. 54, 2393‐2404. Bass, D., Moreira, D., López‐García, P., Polet, S., Chao, E.E., von der Heyden, S., Pawlowski, J., Cavalier‐Smith, T., 2005. Polyubiquitin insertions and the phylogeny of Cercozoa and Rhizaria. Protist 156, 149‐161. Bass, D., Richards, T.A., Matthai, L., Marsh, V., Cavalier‐Smith, T., 2007. DNA evidence an for global dispersal and probable endemicity of Protozoa. BMC Evol. Biol. 7, 162. Bass, D., Chao, E.E., Nikolaev, S., Yabuki, A., Ishida, K., Berney, C., Pakzad, U., Wylezich, C., Cavalier‐Smith, T., 2009. Phylogeny of novel naked filose and reticulose M Cercozoa: Granofilosea cl. n. and Proteomyxidea revised. Protist 160, 75‐109. Beech, P.L., Moestrup, Ø., 1986. Light and electron microscopical observations on the ed heterotrophic protist Thaumatomastix salina comb. nov. (syn. Chrysosphaerella salina) and its allies. Nordic J. Bot. 6, 865‐877. Bowden, R.D., Davidson, E., Savage, K., Arabia, C., Steudler, P., 2004. Chronic nitrogen pt additions reduce total soil respiration and microbial respiration in temperate ce forest soils at the Harvard Forest. Forest Ecol. Man. 196, 43‐56. Brabender, M., Kiss, A.K., Domonell, A., Nitsche, F., Arndt, H., 2012. Phylogenetic and morphological diversity of novel soil cercomonad species with a description of Ac two new genera (Nucleocercomonas and Metabolomonas). Protist 163, 495‐528. Caisová, L., Marin, B., Melkonian, M., 2011. A close‐up view on ITS2 evolution and speciation ‐ a case study in the Ulvophyceae (Chlorophyta, Viridiplantae). BMC Evol. Biol. 11, 262. Cavalier‐Smith, T., 1993. The protozoan phylum Opalozoa. J. Euk. Microbiol. 40, 609‐ 615. Cavalier‐Smith, T., 2013. Early evolution of eukaryote feeding modes, cell structural diversity, and classification of the protozoan phyla Loukozoa, Sulcozoa, and Choanozoa. Eur. J. Protistol. 49, 115‐178. Page 63 of 92 Cavalier‐Smith, T., Chao, E.E., 2003. Phylogeny and classification of phylum Cercozoa 64 (Protozoa). Protist 154, 341‐358. Cavalier‐Smith, T., Chao, E.E., 2012. Oxnerella micra sp. n. (Oxnerellidae fam. n.), a tiny naked centrohelid, and the diversity and evolution of Heliozoa. Protist 163, 574‐ t 601. us cr ip Cavalier‐Smith, T., Karpov, S.A., 2012. Paracercomonas kinetid ultrastructure, origins of the body plan of Cercomonadida, and cytoskeleton evolution in Cercozoa. Protist 163, 47‐75. Cavalier‐Smith, T., Scoble, J.M., 2013. Phylogeny of Heterokonta: Incisomonas marina, a uniciliate gliding opalozoan related to Solenicola (Nanomonadea), and evidence that Actinophryida evolved from raphidophytes. Eur. J. Protistol. 49, 328‐353. an Cavalier‐Smith, T., Lewis, R., Chao, E.E., Oates, B., Bass, D., 2009. Helkesimastix marina n. sp. (Cercozoa: Sainouroidea superfam. n.) a gliding zooflagellate lineage of M novel ultrastructure and unique ciliary pattern. Protist 160, 452‐479. Chantangsi, C., Hoppenrath, M., Leander, B., 2010. Evolutionary relationships among marine cercozoans as inferred from combined SSU and LSU rDNA sequences and ed polyubiquitin insertions. Mol. Phylogen. Evol. 57, 518‐527. Chantangsi, C., Leander, B.S., 2010. An SSU rDNA barcoding approach to the diversity pt of marine interstitial cercozoans, including descriptions of four novel genera and nine novel species. Int. J. Syst. Evol. Microbiol. 60, 1962‐1977. ce Cohen, P.A., Knoll, A.H., 2011. Scale microfossils from the mid‐Proterozoic Fifteenmile Group, Yukon Territory. J. Paleontol. 86, 775‐800. Ac Coleman, A., 2009. Is there a molecular key to the level of "biological species" in eukaryotes? A DNA guide. Mol. Phylogen. Evol. 50, 197‐203. Conrad, W., 1920. Sur un flagellé nouveau à trichocystes, Reckertia sagittifera, n. g., n. sp. Bull. Acad. Roy. Sci. Belg. 6, 541– 555. Dopheide, A., Lear, G., Stott, R., Lewis, G., 2009. Relative diversity and community structure of ciliates in stream biofilms according to molecular and microscopy methods. Appl. Environ. Microbiol. 75, 5261‐5272. Glücksman, E., Snell, E.A., Berney, C., Bass, D., Cavalier‐Smith, T., 2011. The novel marine gliding zooflagellate genus Mantamonas (Mantamonadida ord. n.: Apusozoa). Protist 162, 207‐221. Page 64 of 92 65 Gooday, A. J., Esteban, G. F., Clarke, K. J. 2006. Organic and silicious protistan scales in North East Atlantic abyssal sediments. J. Mar. Biol. Ass. UK. 86, 679‐688. Haska, C.L., Yarish, C., Kraemer, G., Blaschik, N., Whitlatch, R., Zhang, H., Lin, S., 2012. Bait worm packaging as a potential vector of invasive species. Biological t Invasions 14, 481‐493. us cr ip Hawkes, C.V., Kivlin, S.N., Rocca, J.D., Huguet, V., Thomsen, M.A., Suttle, K.B., 2011. Fungal community responses to precipitation. Global Change Biol. 17, 1637‐1645. Heger, T.J., Mitchell, E.A., Todorov, M., Golemansky, V., Lara, E., Leander, B.S., Pawlowski, J., 2010. Molecular phylogeny of euglyphid testate amoebae (Cercozoa: Euglyphida) suggests transitions between marine supralittoral and freshwater/terrestrial environments are infrequent. Mol. Phylogenet. Evol. 55, an 113‐122. Hess, S., Sausen, N., Melkonian, M., 2012. Shedding light on vampires: the phylogeny M of vampyrellid amoebae revisited. PLoS One 7, e31165. Hess, S., Melkonian, M., 2013. The mystery of clade X: Orciraptor gen. nov. and Viridiraptor gen. nov. are highly specialised, algivorous amoeboflagellates ed (Glissomonadida, Cercozoa). Protist 164, 706‐747. Howe, A.T., Bass, D., Scoble, J.M., Lewis, R., Vickerman, K., Arndt, H., Cavalier‐Smith, T., pt 2011a. Novel cultured protists identify deep‐branching environmental DNA clades of Cercozoa: new genera Tremula, Micrometopion, Minimassisteria, ce Nudifila, Peregrinia. Protist 162, 332‐372. Howe, A.T., Bass, D., Chao, E.E., Cavalier‐Smith, T., 2011b. New genera, species and Ac improved phylogeny of Glissomonadida (Cercozoa). Protist 162, 710‐722. Karpov, S.A., 2011. Order Thaumatomonadida. In (Ed. Karpov, S.A.) Protista III. Handbook on Zoology. KMK Scientific Press, Saint Petersburg, Moscow. 402‐428 (in Russian). Karpov, S.A., Bass, D., Mylnikov, A.P., Cavalier‐Smith, T., 2006. Molecular phylogeny of Cercomonadidae and kinetid patterns of Cercomonas and Eocercomonas gen. nov. (Cercomonadida, Cercozoa). Protist 157, 125‐158. Katoh, K., Asimenos, G., Toh, H., 2009. Multiple alignment of DNA sequences with MAFFT. In: Posada, D., (Ed., Bioinformatics for DNA sequence analysis; Methods in Molecular Biology 537. New York, Humana Press. pp. 39–64. Page 65 of 92 66 Koetschan, C., Forster, F., Keller, A., Schleicher, T., Ruderisch, B., Schwarz, R., Muller, T., Wolf, M., Schultz, J., 2009. The ITS2 Database III‐‐sequences and structures for phylogeny. Nucleic Acids Res. 38, doi: 10.1093/nar/gkp1966. Larsen, J., Patterson, D.J., 1990. Some flagellates (Protista) from tropical marine us cr ip Lartillot, N., Philippe, H., 2004. A Bayesian mixture model for across‐site t sediments. J. Nat. Hist. 24, 801‐937. heterogeneities in the amino‐acid replacement process. Mol. Biol. Evol. 21, 1095‐ 1109. Lartillot, N., Brinkmann, H., Philippe, H., 2007. Suppression of long‐branch attraction artefacts in the animal phylogeny using a site‐heterogeneous model. BMC Evol. Biol. 7 Suppl 1, S4.\ an Lartillot, N., Philippe, H., 2008. Improvement of molecular phylogenetic inference and the phylogeny of Bilateria. Philos. Trans. R. Soc. Lond. B 363, 1463‐1472. M Lauterborn, R., 1896. Diagnosen neuer Protozoen aus dem Gebiet des Oberrheins. Zool. Anz. 19, 14‐18. Lauterborn, R., 1899. Protozoen‐Studien IV. Theil. Flagellaten aus dem Gebiete des ed Oberrheins. Zeitschr. wiss. Zool. 65, 369‐391 + Taf XVII‐XVIII. Lee, W.J., Patterson, D.J., 2000. Heterotrophic flagellates (Protista) from marine pt sediments of Botany Bay, Australia. J. Nat. History 34, 483‐562. Lesaulnier, C., Papamichail, D., McCorkle, S., Ollivier, B., Skiena, S., Taghavi, S., Zak, D., ce van der Lelie, D., 2008. Elevated atmospheric CO2 affects soil microbial diversity associated with trembling aspen. Environ. Microbiol. 10, 926‐941. Ac Majaneva, M., Rintala, J.‐M., Piisilä, M., Fewer, D.P., Blomster, J., 2011. Comparison of wintertime eukaryotic community from sea ice and open water in the Baltic Sea, based on sequencing of the 18S rRNA gene. Polar Biology 35, doi: 10.1007/s00300‐00011‐01132‐00309. Mikrjukov, K.A., 2002. Thaumatomastix tauryanini sp. n. (Protista, Sarcomonadea, Thaumatomonadida) is a new species of phagotrophic amoebo‐flagellate from the White Sea sediments. Zool. Zhurn. 81, 261‐265. Moestrup, Ø., 1982. Flagellar structure in algae: a review with new observations particularly on the Chrysophyceae, Phaeophyceae (Fucophyceae), Euglenophyceae, and Reckertia, Phycologia 21, 427–528. Page 66 of 92 67 Moldrup, M., Moestrup, O., Hansen, P.J., 2013. Loss of phototaxis and degeneration of an eyespot in long‐term algal cultures: evidence from ultrastructure and behaviour in the dinoflagellate Kryptoperidinium foliaceum. J. Eukaryot. Microbiol. 60, 327‐334. t Mylnikov, A.P., Karpov, S.A., 1993. A new representative of colourless flagellates us cr ip Thaumatomonas seravini sp. n. (Thaumatomonadida, Protista). Zool. Zhurn. 72, 5‐10 (in Russian). Mylnikov, A.P., Mylnikov, A.A., 2003. The new amoeboid flagellate Thaumatomonas zhukovi (Thaumatomonadida, Protozoa). Zool. Zhurn. (in Russian). 82, 1411‐1417. Mylnikov, A.P., Mylnikova, Z.M., 2012. Ultrastructure of the amoeboid flagellate Thaumatomonas zhukovi Mylnikov et Mylnikov (Thaumatomonadida (Shirkina) an Karpov, 1990). Inland Water Biology. 5, 29‐35. Nicholls, K.H., 2012a. Zoelucasa sablensis n. gen. et n. sp. (Cercozoa, incertae sedis), a M new scale‐covered flagellate from marine sandy shores. Acta Protozool. 51, 113– 117. Nicholls, K.H., 2012b. New and little‐known marine and freshwater species of the ed silica‐scaled genera Thaumatomastix and Reckertia (Cercozoa: Thaumatomonadida). J. Mar. Biol. Ass. UK. doi:10.1017/S0025315412001373. pt Ota, S., Eikrem, W., Edvardsen, B., 2012. Ultrastructure and molecular phylogeny of thaumatomonads (Cercozoa) with emphasis on Thaumatomastix salina from ce Oslofjorden, Norway. Protist 163, 560‐573. Philippe, H., Brinkmann, H., Copley, R.R., Moroz, L.L., Nakano, H., Poustka, A.J., Ac Wallberg, A., Peterson, K.J., Telford, M.J., 2011. Acoelomorph flatworms are deuterostomes related to Xenoturbella. Nature 470, 255‐258. Roure, B., Baurain, D., Philippe, H., 2013. Impact of missing data on phylogenies inferred from empirical phylogenomic data sets. Mol. Biol. Evol. 30, 197‐214. Shiratori, T., Yabuki, A., Ishida, K., 2012. Esquamula lacrimiformis n. g., n. sp., a new member of thaumatomonads that lacks siliceous scales. J. Euk. Microbiol. 59, 527‐536. Stamatakis, A., 2006. RAxML‐VI‐HPC: maximum likelihood‐based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688‐2690. Page 67 of 92 68 Stoeck, T., Bruemmer, F., Foissner, W., 2007. Evidence for local ciliate endemism in an alpine anoxic lake. Microb. Ecol. 54, 478‐486. Stoeck, T., Epstein, S., 2003. Novel eukaryotic lineages inferred from small‐subunit rRNA analyses of oxygen‐depleted marine environments. Appl. Environ. t Microbiol. 69, 2657‐2663. us cr ip Takada, H.Y., Morimoto, S., 2010. Soil clone library analyses to evaluate specificity and selectivity of PCR primers targeting fungal 18S rDNA for denaturing‐gradient gel electrophoresis (DGGE). Microbes Envir. 25, 281‐287. Takahashi, E., Hara, S., 1984. Two new marine species of Chrysosphaerella (Chrysophyceae) with reinvestigation of C. salina. Phycologia 23, 103‐109. Takishita, K., Yubuki, N., Kakizoe, N., Inagaki, Y., Maruyama, T., 2007. Diversity of an microbial eukaryotes in sediment at a deep‐sea methane cold seep: surveys of ribosomal DNA libraries from raw sediment samples and two enrichment M cultures. Extremophiles 11, 563‐576. Thaler, M., Lovejoy, C., 2012. Distribution and diversity of a protist predator 291‐299. ed Cryothecomonas (Cercozoa) in Arctic marine waters. J. Eukaryot. Microbiol. 59, Thomsen, H.A., Hällfors, G., Hällfors, S., Ikävalko, J., 1993. New observations on the pt heterotrophic protist genus Thaumatomastix (Thaumatomastigaceae, Protista incertae sedis), with particular emphasis on material from the Baltic Sea. Ann. ce Bot. Fennici 30, 87‐108. Thomsen, H.A., Ikävalko, J., 1997. Species of Thaumatomastix (Thaumatomastigidae, Ac Protista incertae sedis) from the Arctic sea ice biota (North‐East Water Polynya, N. E. Greenland). J. Mar. Systems 10, 263‐277. Thomsen, H.A., Kosman, C., Ikävalko, J., 1995. Three new species of Thaumatomastix (Thaumatomastigidae, protista incertae sedis), a ubiquitous genus from the Antarctic Ice Biota. Eur. J. Protistol. 31, 174‐181. Tong, S.M., 1997. Heterotrophic flagellates and other protists from Southampton Water, UK. Ophelia 47, 71‐131. Vørs, N., 1992a, Heterotrofe Protister (ekskl. dinoflagellater, loricabaerende choanoflagellater og ciliater), in Thomsen, H.A., ed., Plankton i de indre danske Page 68 of 92 69 farvande. Analyse af forekomsten af alger og heterotrofe protister (ekskl. ciliater) i Kattegat, Volume 11: Copenhagen, Miljøministeriet Miljøstyrelsen pp. 195‐250. Vørs, N., 1992b. Heterotrophic amoebae, flagellates and Heliozoa from the Tvärminne area, Gulf of Finland, in 1988‐1990. Ophelia 36, 1‐109. t Waldrop, M.P., Zak, D.R., Blackwood, C.B., Curtis, C.D., Tilman, D., 2006. Resource us cr ip availability controls fungal diversity across a plant diversity gradient. Ecology Letters 9, 1127‐1135. Wujek, D.E., Pershon, L.E., Kadiri, M.O., 2008. Description of a new freshwater species of Thaumatomastix (Protista, Thaumatomonadida) from Nigeria, West Africa. Tropical Freshwater Biol. 17, 13‐20. Wolf, M., Chen, S., Song, J., Ankenbrand, M., Müller, T., 2013. Compensatory base an changes in ITS2 secondary structures correlate with the biological species concept despite intragenomic variability in ITS2 sequences – a proof of concept. M PLoS ONE, 8(6): e66726. doi:10.1371/journal.pone.0066726 Wylezich, C., Mylnikov, A.P., Weitere, M., Arndt, H., 2007. Distribution and phylogenetic relationships of freshwater thaumatomonads with a description of 347‐357. ed the new species Thaumatomonas coloniensis n. sp. J. Eukaryot. Microbiol. 54, pt Yabuki, A., Ishida, K., 2011. Mataza hastifera n. g., n. sp.: a possible new lineage in the Thecofilosea (Cercozoa). J. Eukaryot. Microbiol. 58, 94‐102. ce Yoon, H.S., Grant, J., Tekle, Y.I., Wu, M., Chaon, B.C., Cole, J.C., Logsdon, J.M. Jr.,Patterson, D.J., Bhattacharya, D., Katz, L.A., 2008. Broadly sampled multigene Ac trees of eukaryotes. BMC Evol. Biol. 8: 14, doi:10.1186/1471‐2148‐8‐14. Zolotarev V.A., Mylnikova Z.M., Mylnikov A.P., 2011. The ultrathin structure of amoeboid flagellate Thaumatomastix sp. (Thaumatomonadida (Shirkina) Karpov, 1990), Inland Water Biology, Vol. 4, No. 3, pp. 287‐292. Zuker, M., 2003. Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res. 31, 3406‐3415. Page 69 of 92 an us cr i 70 Table 1. Thaumatomonas and Allas species cell and scale measurements: type strains in bold; lines separate the main clades. Top left of table shows key. UPL1Be2 Ts z. zhukovi GMBGL1 Ts z. zhukovi ed CW Cell width OR PC PC x CL oSL bt/tt oval Scale Length oSW bt/tt oval Scale Width 6.6 - 11.6 4.5 - 5.5 ~ 1.8 15-18 ~1.6 2.3 bt 0.6 - 0.75 bt 0.3 - 0.4 7 1.5 14.3 1.3 0.61/0.55 10.4 Observations triangle scales? lateral holes? / yes 3 - 5 holes 0.34/24 bt 1.8 tt 2.3 yes 3 - 8 holes ovalnes s ratio 7.5 1.7 19.4 1.5 0.50/0.43 0.31/0.21 bt 1.6 tt 2.1 not observed 5 - 7 holes 10.6 8.2 1.8 14.1 1.3 0.66/0.56 0.36/0.22 bt 1.8 tt 2.8 yes 3 - 8 holes yes 2 - 4 holes CCL4B Ts z. zhukovi 10.2 7.7 1.4 13.1 1.2 0.57/0.47 0.32/0.20 bt 1.8 tt 2.5 GMKL4 Ts z. zhukovi 13.1 8.2 1.6 18.1 1.4 0.51/0.44 0.30/0.20 bt 1.6 tt 2.2 HANTSF8 Ts z. saxoni 10.3 7.8 1.3 11.1 1.1 0.63/0.56 0.30/0.22 bt 1.9 tt 2.6 NC02 Ts z. carolinensis 12.2 6.8 1.8 14.4 1.2 0.54/0.44 0.32/0.18 bt 1.7 tt 2.4 yes 2 - 5 holes NC03 Ts z. paracarolinensis 10.7 6.8 1.6 14 1.3 0.54/0.44 0.32/0.18 bt 1.7 tt 2.4 yes 3 - 5 holes 9.9 - 16.5 6.6 - 11.6 ~ 1.5 20 - 33 2 Triangle 0.7 none none yes ~ 5 holes Ts seravini seravini (T-2) (Mynikov and Karpov 1993) NOTES OR 12.5 Ac UPL1Bf2 Ts z. zhukovi Oval scales CL Cell length pt Ts zhukovi zhukovi (T-3) (Mylinkov and Mylnikov 2003) Electron microscopic measurements μm Light microscopic live-cell measurements μm ce / no data ~ approximately ? unknown PC posterior cilium bt bottom tier tt top tier M ______________________________________________________________________________________________________________________________ not observed ~yes (trianglelike) 3 - 8 holes 2 - 10 holes Page 70 of 92 triangular, oval, and intermediate scales round circle apertures at either end of scale form broad struts 3 - 5 solid struts on oval scales (no 18S rDNA sequence) round circle apertures at either end of scale form broad struts round circle apertures at either end of scale form broad struts 3 - 5 solid struts (scale) round circle apertures at either end of scale form broad struts round circle apertures at either end of scale form broad struts round circle apertures at either end of scale form broad struts round circle apertures at either end of scale form broad struts 7.5 1.4 14.1 4e Ts s. varisquama 11.9 9.5 1.3 14.7 IVY18a Ts sp. 10.7 7.4 1.5 8.3 - 14.9 5 - 8.3 ~ 1.7 Ts constricta CH3 12.2 7.4 BZ1 Ts contricta 10.8 9.1 - 14.6 Ts coloniensis 'vancouveri' (3108W2) PML3A Ts coloniensis MLTB12 Ts coloniensis Ts oxoniensis (Hinksey) GMPL1 Ts oxoniensis none none yes 3 - 5 holes 1.2 Triangle 0.48 oval 0.56/0.49 oval 0.31/0.21 bt 1.9 tt 2.4 yes 3 - 4 holes 1.2 / / / / / 14.4 2 bt ~ 0.7 bt ~ 0.35 ? no none 15.5 1.3 0.57/0.44 0.28/0.08 bt 2.1 tt 5.3 no none scale top tier is very narrow with no lateral holes scale top tier is very narrow with no lateral holes ed 13.1 14.9 1.4 0.57/0.47 0.27/0.08 bt 2.1 tt 6.4 no none 3 - 5.8 ~2 16 - 18 ~ 1.2 bt 0.50-0.65 bt 0.35-0.45 ~bt 1.5 no ~5-7 holes 1.3 13.3 1.1 / / / / / 8.2 - 11.8 round circle apertures at either end of scale form broad struts oval scales; 3 - 5 solid struts. Scales difficult to measure - broken molecularly closet to T. seravini but with an 18S rDNA nucleotide difference 4 - 5 solid struts (scales) 8 (from fixed) ? ? ~ 9.6 1.2 / / / no "several" 4 solid struts (data from Howe et al, 2011a). Small measurement due to shrinkage of fixed cell? 11.8 7 1.7 16.7 1.5 0.52/0.45 0.30/0.22 bt 1.8 tt 2.1 no 5 - 7 holes 3 - 5 solid struts (scales) Ac Ts coloniensis-like ('lauterborni' from K. Vickerman) 9.9 - 14.9 Triangle 0.5 1.5 pt Ts coloniensis (HFCC59) (Wlyezich et al 2007) 1.7 1.4 7.5 ce Ts lauterborni (T-1) (Mylnikov and Karpov, 1993) an us cr i 10.3 M CH8 Ts s. seravini 71 13.4 7.4 1.8 17.6 1.3 0.53/0.46 0.29/0.21 bt 1.8 tt 2.2 yes 4 - 6 holes Broad struts at either end of oval scale and other scales with smaller solid struts. Oval and triangular scales observed 9.5 ~6 1.6 18.5 2 / / / no 4 holes 4 - 5 solid struts (scales) 13 8.1 1.6 17.5 1.4 0.50/0.40 0.27/0.20 bt 1.9 tt 2.2 no 4 - 6 holes 3 - 5 solid struts (scales) Page 71 of 92 7.4 1.7 15.5 14.2 11 1.3 15.9 no data / / / 13 - 20 ? ? Allas multipora GSPB9 12.2 9.6 1.3 Allas media BKARH9 11.4 8.9 1.3 0.51/0.44 0.28/0.21 1.2 0.57/0.49 0.31/0.23 ~1.3 0.55/0.47 0.33/0.18 ~1.1 ~ 0.5 bt ~ 0.35 13.8 1.1 0.61/0.53 17.5 1.5 0.58/- ed 13 - 20 bt 1.8 tt 2.2 bt 1.9 tt 2.7 bt 1.7 tt 2.6 no 5 - 7 holes 3 - 5 solid struts (scales) no 4 - 8 holes 3 - 5 solid struts (scales) no ~5-7 3 - 4 solid struts (scales) ? lateral strut triangular up to 4 distinct large lateral holes on scale upper tier 0.4/0.25 bt 1.7 tt 2.2 lateral strut triangular up to 8 bt 0.47 bt 1.2 lateral strut triangular up to 10 Ac ce pt Allas aff. diplophysa (ATCC50635 Howe et al, 2011a) 1.3 M 12.9 an us cr i EP3 Ts oxoniensis Ts solis BZ8 Ts solis CH9 72 Page 72 of 92 bottom tier of scale notably wider than top tier scales appear in between oval to triangular in shape an us cr i 73 Table 2. Reckertia, Ovaloplaca, Scutellomonas and Thaumatospina cell and scale measurements: Top left of table shows key. CL 11.4 ed Reckertia filosa (JJP-2003 Howe et al, 2011a) Reckertia gemma VWB1 THAS Light microscope live-cell measurements μm Electron microscopic measurements - μm Spine scale Plate scale NOTES CW OR PC PC x CL Spine length Base width Oval Triangular ? ? 17 - 20 1.8 - 2 / / / ~ 0.55 oval AC scales 0.33 - 0.37 x 0.2 - 0.22 pt / no data ~ approximately oSW oval Scale Width oSL oval Scale Length OR ovalness ratio M _____________________________________________________________________________________________________________________________________________________ 9.7 1.2 19.1 1.6 / / / ~ 0.45 each side oval AC scales 0.25 - 0.36 x 0.15 - 0.19 10.7 7.5 1.4 ~ 15 ~ 1.5 ~ 5.4 / oSL ~ 1.3 oSW ~ 0.9 none scales measured from Figure 2. Ota et al., 2012 11.1 10. 5 1.1 11 1 5.18 / oSL 1.3 oSW 0.7 none one cell measured 10.6 8.5 1.3 15.2 1.4 no spine scale no spine scale 0.77 x 0.53 none 11.9 6.7 1.8 17.7 1.6 4 0.64 none 1.2 Thaumatospina mexica TxMEX / / / / / 4.1 0.58 none 1.2 Thaumatospina arabica TxDUBAI 14.9 10 1.5 20.8 1.4 / / none yes ce 11.6 Ovaloplaca salina OTA et al, 2012 Ac Ovaloplaca yabukii YPF609 Scutellomonas patella CA05 Thaumatospina vancouveri VWB1C Page 73 of 92 observed in LM but not measured (large 18S rDNA insert) 74 Table 3. Provenance of Thaumatomonas isolates __________________________________________________________________ Spe cies Pr ove n an ce T‐1 Thaumatomonas lauterborni Borok, Russia BZ1 Thaumatomonas constricta ?, Brazil CH3 Thaumatomonas constricta Yangshuo, China T‐2 Thaumatomonas s. seravini Borok, Russia 4E Thaumatomonas s. varisquama Oxford, United Kingdom T‐3 Thaumatomonas z. zhukovi Borok, Russia GMKL4 Thaumatomonas z. zhukovi Staffordshire, United Kingdom (2009) UPL1Be2 Thaumatomonas z. zhukovi Oxford, United Kingdom (2009) UPL1Bf2 Thaumatomonas z. zhukovi Oxford, United Kingdom (2009) CCL4B Thaumatomonas z. zhukovi Oxford, United Kingdom (2009) GMBGL1 Thaumatomonas z. zhukovi Oxford, United Kingdom (2009) HANTSF8 Thaumatomonas z. saxoni NC02 Thaumatomonas z. carolinensis NC03 Thaumatomonas z. paracarolinensis Hinksey Thaumatomonas oxoniensis GMPL1 Thaumatomonas oxoniensis EP3 Thaumatomonas oxoniensis Exmoor, United Kingdom (2009) HFCC59 Thaumatomonas coloniensis Cologne, Germany PML3A Thaumatomonas coloniensis Oxford, United Kingdom (2009) MLTB12 Thaumatomonas coloniensis ?, Malta (2002) 3108W2 Thaumatomonas coloniensis Vancouver, Canada CH9 Thaumatomonas solis Yunnan, China us cr ip an Hampshire, United Kingdom (2000) North Carolina, USA ed M North Carolina, USA Oxford, United Kingdom Staffordshire, United Kingdom Ac ce pt t St r a in Page 74 of 92 75 Figure legends Figure 1. PhyloBayes CAT‐GTR gamma tree for 273 Monadofilosa 18S rDNA sequences (1785 nucleotide positions) with environmental sequences in bold. The tree is across two pages: Fig. t 1.1, Ventrifilosa: Imbricatea and Thecofilosea; Fig. 1.2, Sarcomonadea. Arrows indicate new us cr ip species from this study. Asterisks indicate new Thaumatomonadidae species type strains from this study. Environmental clades mentioned in text: R1 environmental marine sister to Reckertia; MF1 environmental Marine Flagellate clade 1; NC1, NC3, and NC4, from Bass and Cavalier‐Smith (2004); eVentri Ventricleftida; eEbriida sister to Ebriida; eSarcomonad. Paracercomonads labeled according to Brabender et al. (2012). FW shows the base of the major freshwater clade within Thaumatomonadidae; deeper branches in this family are an exclusively marine. Support values are PhyloBayes posterior probabilities (left), RAxML GTR gamma bootstrap percentages for 1000 pseudoreplicates (middle), MrBayes posterior probabilities (right). Black dots indicate maximal support for all, i.e. 1/100/1. The tree is rooted M between cercomonads and glissomonad/pansomonads in accord with the cercozoan‐wide tree of Howe et al. (2011a). The robust clade lying between Pansomonadida and Glissomonadida, which includes Viridiraptoridae (Hess and Melkonian 2013) and clade Y are here assigned to ed Pseudosporida rather than Glissomonadida for reasons explained in the discussion. Figure 2. MrBayes GTR gamma covarion tree for 35 species of Thaumatomonadidae based on pt ITS1, 5.8S, ITS2 rDNA sequences (1561 nucleotide positions) including ~400 nucleotides of the end of 18S rDNA and the first ~300 nucleotides of 28S rDNA. Sketches of scales from each ce strain are shown; for T. z. zhukovi T‐3 parentheses embrace additional scale types observed in the original publication (Mylnikov and Mylnikov 2003); for MLTB12 parentheses embrace Ac scales observed from one bad preparation unable to confirm which type. Asterisks indicate species type strains, and arrows indicate new subspecies. Smaller anterior cilium scales for Reckertia are shown to the right of the body‐scales. Some genera and species from Figure 1 have no ITS sequences available. Support values are MrBayes posterior probabilities (left) and RAxML bootstrap percentages for 1000 pseudoreplicates (right). Black dots indicate maximal support for both, i.e. 1/100. Figure 3. Secondary structure of ITS2 rDNA folds. A. Examples of Thaumatomonas species: Ts constricta, Ts seravini varisquama, Ts z. zhukovi, Ts oxoniensis, Ts s. seravini. Enlarged loop of Ts seravini varisquama labeled on the third helix. B. Reckertia gemma. C. Thaumatospina: Ta arabica and Ta mexica. D. Allas: A. media and A. multipora. E. Scutellomonas patella. All Page 75 of 92 76 secondary structures have similar folds and show four clear helices (labeled on Thaumatomonas seravini varisquama). Figure 4. A‐B, Thaumatomonas zhukovi saxoni (HANTSF8). C‐D, Thaumatomonas seravini varisquama (4e). E, Thaumatomonas sp. (IVY18a). F‐G, Thaumatomonas seravini seravini (T‐2). us cr ip t A. Body‐scales of Ts zhukovi saxoni (HANTSF8); less common triangular‐like (*), bowed upper tier and end strut (arrow). B. DIC of live Ts z. saxoni (HANTSF8); i – ii, one gliding cell in two focal planes (two‐way arrow), i, anterior cilium (AC) and, ii, posterior cilium (PC), iii – iv, one feeding cell shown in two focal planes (two‐way arrow), iii, cilium pit (PIT) and, iv, pseudopodia. Cell with ventral groove (double arrowhead) and anterior cilium (AC). Tilted side view of cell (*) with anterior cilium (AC). C. DIC of live Ts s. varisquama (4e); i – ii, gliding partially divided cell (%) shown in two focal planes (two‐way arrow), with posterior cilium (PC), iii, two feeding an cells with pseudopodia, broad lamellipodia (*), iv – v, two gliding cells with broad posterior end and anterior cilium (AC), vi, three feeding cells joined via pseudopodia each with posterior M cilium (PC). Gliding dividing cell (%) with anterior cilium (AC). Gliding cell with anterior labeled (ant.). D. Body‐scales of Ts s. varisquama (4e) showing five narrow end struts on oval scales and single broad vertex struts on triangular scales. E. DIC of live Ts sp. (IVY18a; 18S groups ed closely with Ts seravini); i – ii, double feeding cell on two focal planes (two‐way arrow) with branching pseudopodia (double arrowhead), iii, single cell with pointed posterior, anterior indicated (ant.). F. Side view of triangular Ts s. seravini (T‐2) body‐scale with corner strut pt (arrow). G. Body‐scales of Ts s. seravini (T‐2). Scale bars: B, C and E, 10 μm. A, D, F and G, 0.5 ce μm. Figure 5. A‐B, Thaumatomonas constricta (BZ1). C‐D, Thaumatomonas solis (BZ8 type). E‐G, Ac Thaumatomonas constricta (CH3 type). H, Thaumatomonas solis (CH9). A. DIC of live Ts constricta (BZ1); i – ii, gliding cells showing posterior cilium (PC), ii, with two posterior cilia (**), iii, two feeding cells joined via long thin pseudopodium (double arrowhead), iv, gliding cells with flared end of posterior cilium (*), v, two gliding cells with anterior cilium (AC) and one with ventral groove (arrow). B. Ts constricta (BZ1) body‐scales, side view of end strut (double arrowhead). C. DIC of Ts solis (BZ8) cells; oblong cell with two sets of cilia (**), i, iii, and iv – v, gliding cells with broad posterior ends (Post.) and visible anterior cilium (AC), ii, feeding cell, iii, ventral groove (arrow). D. Body‐scales of Ts solis (BZ8 type) aberrant scale (*); arrows indicate discrete narrow end struts. E. DIC of live Ts constricta (CH3); i – ii, two feeding cells with thick branching pseudopodia (double arrowheads), small cell (sm.), and round cyst, iii – iv, gliding cells, anterior end (ant.), ventral groove (arrow). F. TEM of Ts constricta (CH3) Page 76 of 92 77 body‐scales, side view with end struts (double arrowhead). G. TEM of Ts constricta (CH3) cell with posterior and anterior cilia (PC, AC). H. Body‐scales of Ts solis (CH9). Scale bars: A, C and E 10 μm. B, D, F and H 0.5 μm, G 2 μm. Figure 6. A‐B, Thaumatospina vancouveri (VWB1C). C‐F, Thaumatospina mexica (TxMex). us cr ip t A. DIC of live Ta vancouveri (VWB1C); i – ii, same cell (two‐way arrow) in two focal planes showing a broad pseudopodium (double arrowhead), posterior and anterior cilia (PC, AC), prominent nuclei (*). Close‐up of flared end of posterior cilium (arrow in bordered window). iii – iv, two gliding cells with prominent AC and ventral groove (arrow). Aggregate refractile mass of cells (**). B. TEM of Ta vancouveri (VWB1C) body‐scales, some plate scales with light inner triangle shape (arrow). Three normal spine‐scales and one aberrant with highly curved spine and elongate base (*). Close‐up of trifurcate tip (window). C‐G, TEM of Ta mexica (Mex), C. an Body‐scales, aberrant spine‐scale with two spines (*); punctate markings on the margin of the upper tier – magnified in window with asterisks (**). D. Cell with no spine‐scales. E. Spine‐ scale; flared and pointed tip (left), upper flange and base plate (right). F. Whole cell with spine G 1 μm, D and F 2 μm, E 0.5 μm. M and plate‐scales. G. Triangular plate‐scales, and two spine‐scales. Scale bars: A 10 μm. B, C and ed Figure 7. A, Cowlomonas planata (VWB1B). B‐C, Thaumatospina arabica (Dubai). A. DIC of two live C. planata (VWB1B); i – x, movement series of same cell, vii – viii, ventral pt ridge continuation of the cell’s anterior hood (double arrowhead), x, ventral pseudopodia at posterior end (arrowhead), xi – xiii, movement series of same cell, xi, posterior and anterior ce cilia (AC, PC) coming out of hood. B. DIC of live Ta arabica (Dubai); i – ii, spine‐scales (double arrowhead), i, ii, and v – vi, anterior cilium (AC), iii, iv and viii, posterior cilium (PC), iii – iv, Ac feeding cells with branching pseudopodia (arrowhead), vii, ventral groove (arrow). C. TEM of Ta arabica (Dubai) spine and two‐tiered plate‐scales. Punctate markings on upper tier underside margin of triangular plate scales (*) next to bobbin base of spine‐scale. Scale bar: A and B 10 μm, C unknown scale. Figure 8. A‐D, Scutellomonas patella (CA05). E‐G, Ovaloplaca yabukii (YPF609). H‐J, Reckertia gemma (VWB1Thas). A. DIC of live S. patella (CA05); i–v, series of a cell creeping with active pseudopodia; ii, posterior cilium and anterior orientation of the cell (ant.); vi, a cell with ventral groove (arrow); vii – ix, series of a third cell, gliding, vii shows a ridge at the anterior end (*); ix and x, thinning posterior cilium distal end, which appears acronematic (arrowhead), close‐up (window); x ‐ xi, Page 77 of 92 same cell in different focal plane, xi showing subapical ciliary pit. B. TEM of S. patella (CA05), 78 showing body‐scales and posterior cilium with thin tip. C. TEM of S. patella (CA05) cell with very short second cilium (*), possibly pre‐divisional. D. Body‐scales of S. patella (CA05) showing four main struts (**). E. Phase contrast micrograph of O. yabukii (YPF609) showing posterior cilium (PC) and spine‐scales (double arrowhead). F‐G. SEM of body‐scales of O. us cr ip t yabukii (YPF609); dished oval scales with upper‐tier inner perforations, and ribbed spine‐scales with proximal flange above base‐plate and tripartite basally flared struts linking them. H. Body‐ scales of R. gemma with triangular base‐plate (arrowhead) and inner margin; seen on three body scales (double arrowhead); corner long strut has short bifurcation onto lower margin creating small aperture (arrow). Side view of central, domed upper tier ‘jewel’ (*). I. Cilium scales of R. gemma (VWB1Thas). J. DIC of R. gemma (VWB1Thas); i – ii, four gliding cells showing anterior cilium (AC) and angled apical dent at ciliary pocket; iii, anterior cilia of two an cells; iv – v, (binucleate and tetraciliate) feeding cells on different focal planes (two‐way arrow); v, branching pseudopodia, vi group of gliding cells showing posterior cilia (PC), vii – viii, gliding cells with anterior end indicated (ant.), with large nuclei with dense nucleolus visible M (near ciliary bases). Scale bars: A, E and J 10 μm. B, and G 2 μm. C and F 1 μm, H and I 0.2 μm, D 0.5 μm. ed Figure 9. A‐B, Allas multipora (GSPB9). C‐E, Allas media (BKARH9). A. DIC of live gliding A. multipora (GSPB9); anterior and posterior cilia (AC, PC). i, thin ventral pt groove (arrow), ii and iv – v, cells with pointed and/or ‘ragged’ posterior ends in contrast to iii, with broad rounded posterior B. Body‐scales of A. multipora (GSPB9); side view of broad ce lateral strut (double arrowhead). C. TEM of whole mount A. media (BKARH9) cell, anterior and posterior cilia (PC, AC), PC with thin tip. D. Body‐scales of A. media (BKARH9). Side view of Ac scale with broad strut (double arrowhead). E. DIC of live A. media (BKARH9), i – ii, same polynucleate round mass (two‐way arrow) in different focal planes, ii, with cilium, iii, gliding cell mass with two posterior cilia (PC), iv, ventral groove (double arrowhead) on gliding cell, anterior and posterior cilia (AC, PC), v, complete length of posterior cilium on gliding cell (PC), vi, granulated ventral groove (arrow). Scale bars: A and E 10 μm. B and D, 0.5 μm, C 2 μm. Figure 10. A‐B, Thaumatomonas zhukovi paracarolinensis (NC03). C‐D, Thaumatomonas zhukovi zhukovi (GMKL4). E‐F, Thaumatomonas zhukovi zhukovi (GMBGL1). G‐I, Thaumatomonas zhukovi carolinensis (NC02). A. DIC of Ts z. paracarolinensis (NC03); i, elongate dividing gliding cell (%), ii, small round cell, iii – iv, same gliding cell shown on two focal planes (two‐way arrow) posterior cilium (PC), v – vi, Page 78 of 92 79 same small round feeding cell on two focal planes (two‐way arrow) with pseudopodia (double arrowhead), vi, large feeding cell with branched pseudopodia (double arrowheads). Refractile aggregate of cells of various sizes (*). B. Body‐scales of Ts z. paracarolinensis (NC03). C. DIC of live Ts z. zhukovi (GMKL4); i, feeding cell with pseudopodia (double arrowheads), ii, gliding cell with pointed posterior and visible anterior cilium (AC), iii, gliding cell with two sets of anterior us cr ip t and posterior cilia (AC, PC). D. Body‐scales of Ts z. zhukovi (GMKL4). E. Body‐scales of Ts z. zhukovi (GMBGL1). F. DIC of live Ts z. zhukovi (GMBGL1); i, feeding cell with branching pseudopodia (double arrowhead), ii – iv, gliding cells with visible posterior and anterior cilia (PC, AC), ii – iii, tapered posterior. G. TEM of Ts z. carolinensis (NC02) body‐scales H. TEM of whole Ts z. carolinensis (NC02) cell. I. DIC of live Ts z. carolinensis (NC02) cells; i, small round cell at the right of an oval gliding cell with anterior end indicated (ant.), ii, gliding cell with posterior obtusely angled (flick), anterior cilium indicated on gliding cells (AC), anterior of an gliding cell (ant.), anterior cilium on gliding cells (AC), iii – iv, feeding cells show broad pseudopodia (double arrowhead) and posterior cilium (PC). Scale bars: A, C, F and I 10 μm, B, D, M E and G 0.5 μm, H 2 μm. Figure 11. Hypothesis for scale evolution in Imbricatea. Ancestrally, imbricates probably had unadorned oval single‐tier scales made in Golgi‐derived cisternae like those of Euglyphida. ed Thaumatomonadida alone evolved a novel shaping process by attachment of the Golgi cisternae to the mitochondrial envelope and its invagination to make a second intramembrane pt space for addition of the upper tier. Peregriniidae retained the simple oval shape; Gyromitus has a symmetric latticed wall connecting the lower and upper tiers, whereas Peregrinia ce modified this to make distinct struts positioned at the ends of the scale only. The ancestor of Thaumatomonadidae evolved three symmetric struts making the plate scales triangular and differentiated them into two types: plate scales, and more radically modified spine scales in Ac which the upper tier was extended centrally as a long hollow spine and contracted laterally to make a basal flange connected to the lower tier by three struts; fused distally but splaying out dichotomously at their base. This bobbin‐based spine scale was retained by Thaumatospina and Ovaloplaca but lost independently by Scutellomonas and the clade including Thaumatomonas, Allas, and Reckertia (and putatively Thaumatomastix). Thaumatomastix scales evolved non‐homologous lateral spines without a flange. Within Thaumatomonadidae, oval scales evolved thrice independently: the ancestor of the Scutellomonas/Ovaloplaca clade evolved numerous symmetrically arranged almost invisible struts around the inner oval; Allas made the lower tier only oval and squashed the triangular upper tier laterally; Thaumatomonas made the lower tier oval and the upper tier oblong, but the Ts Page 79 of 92 zhukovi/seravini ancestor initially retained triangular scales also. The Allas/Thaumatomonas 80 ancestor evolved lateral and end/vertex perforations in the bowed upper tier. End struts increased in number independently in the Ts lauterborni/solis ancestor and at least one sublineage of Ts zhukovi. Single tier anterior ciliary scales (Ovaloplaca and Reckertia only) evolved, probably from single tier scale rudiments prior to migration to the mitochondria in us cr ip t the common ancestor of all Thaumatomonadidae except Thaumatospina, and were lost by Ac ce pt ed M an Scutellomonas and the Thaumatomonas/Allas/Thaumatomastix ancestor. Page 80 of 92 Figure1.1 0.2 Thaumatomonas z. zhukovi (T-3) KC243119 Thaumatomonas z. zhukovi ;hW>ϭĨϮͿ<ϮϰϯϭϬϱ Thaumatomonas sp. (‘Allas’ SA) AF411263 Thaumatomonas 1/99/1 Thaumatomonas sp. (SA) AF411260 Thaumatomonas sp. (Ivy18a) KC243106 Thaumatomonas s. seravini;dϱϬϲϯϲͿ;dͲϮͿ&ϰϱϱϳϳϲ 0.98/93/1 Thaumatomonas s. seravini ^WdϮ;dͲϮͿzϰϵϲϬϰϱ Thaumatomonas s. seravini ;dϱϬϲϯϲͿ;dͲϮͿ&ϰϭϭϮϱϵ Allas Ăī͘diplophysa AF411262 Allas media ;Ύ<Z,ϵͿ<ϮϰϯϭϬϴ Allas ůůĂƐŵƵůƟƉŽƌĂ;Ύ'^WϵͿ<ϮϰϯϭϬϳ Thaumatomonas coloniensis ;,&ϱϵͿYϮϭϭϱϵϭ Thaumatomonas coloniensis (3108W2) AF411264 Thaumatomonas sp. U42446 Thaumatomonas ‘coloniensisͲůŝŬĞ͛^WdϭzϰϵϲϬϰϲ Freshwater Ϭ͘ϴϲͬͲͬϬ͘ϱϭ Thaumatomonas oxoniensis CCAP1903/2 HQ121429 sp. (P106) KC243120 Thaumatomonas 1 Thaumatomonas Ϭ͘ϰϳͬϴϭͬϬ͘ϵϳ Thaumatomonas constricta (*CH3) KC243110 Thaumatomonas constricta ;ϭͿ<ϮϰϯϭϬϵ 0.99/94/1 Thaumatomonas lauterborni Wdϯ;dͲϭͿzϰϵϲϬϰϰ Thaumatomonas Ăī͘solis (CH9) KC243112 Thaumatomonas solis ;ΎϴͿ<Ϯϰϯϭϭϭ Thaumatomonas sp. dϱϬϮϱϬ&ϰϭϭϮϲϭ AB534345 Uncultured eukaryote EF023728 Thaumatomonad Amb 18S 1164 Ϭ͘ϱϯͬϱϱͬϬ͘ϳϱ FN690394 Uncultured cercozoan 5-B11 FN690393 Uncultured cercozoan 3b-D9 ‘R1’ sea ice clade FN690392 Uncultured cercozoan 8-32 ZĞĐŬĞƌƟĂŐĞŵŵĂ;ΎstϭdŚĂƐͿ<Ϯϰϯϭϭϳ ZĞĐŬĞƌƟĂĮůŽƐĂ;::WͲϮϬϬϯͿzϮϲϴϬϰϬ Ϭ͘ϱϭͬϰϲͬϬ͘ϳϴ EF100294 Uncultured eukaryote GU385595 Uncultured marine eukaryote ME Euk Fw1 Ovaloplaca salina&Zϴϰϲϭϵϲ Ϭ͘ϰϵͬϱϮͬϬ͘ϴϭ Ovaloplaca Ϭ͘ϱϭͬϰϲͬϬ͘ϳϳ Ovaloplaca yabukii ;ΎzW&ϲϬϵͿ<Ϯϰϯϭϭϴ 0.99/100/1 AY180026 Uncultured cercozoan CCW52 Scutellomonas Scutellomonas patella ;ΎϬϱͿ<Ϯϰϯϭϭϲ Ϭ͘ϱϰͬϱϭͬϬ͘ϴϲ Ϭ͘ϱϲͬϯϴͬϬ͘ϴϭ AY620310 Uncultured cercozoan 9-2.6 AY620317 Uncultured cercozoan A15 Thaumatospina vancouveri;ΎstϭͿ<Ϯϰϯϭϭϰ Thaumatospina mexica ;ΎDĞdžϭͿ<Ϯϰϯϭϭϱ 0.99/90/1 Thaumatospina Ϭ͘ϵϳͬͲͬͲ Thaumatospina arabica (*Dubai) KC243113 Thaumatospina sp. GQ144681 EF023480 Thaumatomonad Amb 18S 1066 0.48/-/WĞƌĞŐƌŝŶŝĂĐůĂǀŝĚĞĨĞƌĞŶƐYϮϭϭϱϵϯ ƐƋƵĂŵƵůĂůĂĐƌŝŵŝĨŽƌŵŝƐϳϬϴ^^hϳϭϰϮϳϬ Ϭ͘ϵϳͬϭϬϬͬϭ FR874390 Uncultured marine picoeukaryote ws 96 &:ϰϭϬϵϭϲDĂƌŝŶĞŇĂŐĞůůĂƚĞϱ ‘MF1 clade 0.61/-/Ϭ͘ϳϱͬϵϰͬϭ Spongomonas solitaria,YϭϮϭϰϯϱ Spongomonas minima AF411281 Spongomonas ƐƉ͘<sͬzϲϮϬϮϱϮ 1/98/1 Euglypha rotunda:ϰϭϴϳϴϮ ƵŐůLJƉŚĂĮůŝĨĞƌĂ:ϰϭϴϳϴϲ Euglypha rotunda ;WϭϱϮϬͿyϳϳϲϵϮ Ϭ͘ϲϱͬͲͬϬ͘ϲϮ 0.13/13/Euglypha rotunda :ϰϭϴϳϴϰ 0.92/-/Trachelocorythion pulchellum:ϰϭϴϳϴϵ Corythio dubium &ϰϱϲϳϱϭ Trinema enchelys:ϰϭϴϳϵϮ Euglypha acanthophora :ϰϭϴϳϴϴ Assulina muscorum :ϰϭϴϳϵϭ 0.99/100/0.99 Ovulinata parva HQ121432 Ovulinata ƐƉ͘;zW&ϱϬϭͿϰϱϯϬϬϮ Paulinella chromatophorayϴϭϴϭϭ Ϭ͘ϴϳͬͲͬͲ Cyphoderia ampulla GU228896 0.48/32/0.94 Cyphoderia ampulla :ϰϭϴϳϵϯ 0.39/-/AY620296 Uncultured FW cercozoan 13-1.8 AY620333 Uncultured cercozoan 7-1.3 AY620334 Uncultured cercozoan 7-1.4 NC3 (Bass &TCS 2004) 0.41/-/0.3/-/AY620308 Uncultured marine cercozoan 9-2.2 NC4 (Bass & TCS 2004) Ϭ͘ϳϰͬϰϳͬϬ͘ϱϯ AB505503 Uncultured eukaryote RM1-SGM46 0.22/-/- 0.39/-/0.32 Pseudopirsonia mucosa :ϱϲϭϭϭϲ ƵƌĂŶƟĐŽƌĚŝƐƋƵĂĚƌŝǀĞƌďĞƌŝƐEU484394 0.99/100/1 0.98/90/0.99 AY620332 Uncultured marine cercozoan 7-5.4 AY620359 Uncultured marine cercozoan 12-4.3 NC1 (Bass &TCS 2004) 0.8/-/0.99 AY620337 Uncultured marine cercozoan 12-4.1 0.43/-/EF023758 ‘Thaumatomonad’ Amb 18S 1199 Ϭ͘ϳϴͬϴϮͬϬ͘ϵϰ EƵĚŝĮůĂƉƌŽĚƵĐƚĂHQ121434 ůĂƵƚƌĂǀŝĂďŝŇĂŐĞůůĂƚĂ&:ϵϭϵϳϳϮ x3 Ϭ͘ϵϳͬϴϱͬϭ 1/99/1 AY620331 Marine cercozoan C5 0.36/21/0.98 Ϭ͘ϳϲͬϯϳͬϬ͘ϴϭ AY620329 Marine cercozoan C1 AY620336 Uncultured marine sediment 7-6.6 NC1 (Bass & TCS2004) ϲϵϱϱϭϵhŶĐƵůƚƵƌĞĚĂŶƚĂƌĐƟĐĞƵŬĂƌLJŽƚĞDWϮͲϮϱ AY620304 Uncultured cercozoan 13-2.8 0.34/-/FN690387 Uncultured cercozoan 5-A10 0.99/98/1 AY620316 Uncultured cercozoan D6 Discomonas retusa FJ824131 AY620347 Uncultured cercozoan 7-5.5 sĞŶƚƌŝĮƐƐƵƌĂĂƌƚŽĐĂƌƉŽŝĚĞĂ&:ϴϮϰϭϮϳ 0.99/100/1 EF526931 Uncultured marine SA2_4G8 AY180018 Uncultured cercozoan CCW29 AY620350 Uncultured cercozoan 12-4.4 sĞŶƚƌŝĮƐƐƵƌĂĨƵƐŝĨŽƌŵŝƐ FJ824128 0.39/42/0.42 0.99/-/EF526932 Uncultured marine SA24H12 EU050976 Uncultured eukaryote SS1 E 01 09 Ϭ͘ϵϵͬϵϳͬϭ FN690383 Uncultured cercozoan 3b-F9 ‘eVentri’ clade AY620349 Uncultured c12-3.6 Ϭ͘ϵϳͬϳϭͬϬ͘ϵϵ EU050975 Uncultured eukaryote SS1 E 01 26 AY180012 Uncultured cercozoan CCW10 EU371397 Uncultured eukaryote NPKS2_52 0.49/-/sĞƌƌƵĐŽŵŽŶĂƐďŝĮĚĂFJ824129 sĞƌƌƵĐŽŵŽŶĂƐůŽŶŐŝĮůĂ FJ824130 AB505573 Uncultured eukaryote RSM-SGM65 Protaspa sp. FJ824124 Protaspa sp. FJ824123 FN690412 Uncultured cercozoan 3c-H7 Ϭ͘ϵϵͬϵϱͬϭ Protaspa longipes&ϮϵϬϱϰϬ Protaspa ƐƉ͘&:ϴϮϰϭϮϱ Protaspa grandis DQ303924 EF024013 Soil clone Amb Ϭ͘ϱϮͬͲͬͲ Rhogostoma minus :ϱϭϰϴϲϳ Rhogostoma sp HQ121436 Ϭ͘ϰϰͬϴϱͬͲ 0.93/-/Rhogostoma scheussleri HQ121430 Ϭ͘ϲϱͬϵϲͬϬ͘ϯϮ FN690411 Uncultured cercozoan 3b-3G HM135090 Uncultured cercozoan STfeb 183 0.94/82/1 AB275093 Uncultured eukaryote CYSGM-10 AB275046 Uncultured eukaryote DSGM-46 0.36/-/JF698753 Uncultured cercozoan MALINA St390 DQ314809 Uncultured marine NOR26.10 0.99/90/1 JN048125 Uncultured eukaryote NB 23A9G ƌLJŽƚŚĞĐŽŵŽŶĂƐĂĞƐƟǀĂ&ϮϵϬϱϯϵ 0.69/31/0.82 FN690365 Uncultured cercozoan 4-F11 FN690408 Uncultured cercozoan 5-H10 FN690404 Uncultured cercozoan 8-28 Ϭ͘ϵϵͬϴϳͬϭ HQ696569 Uncultured clone 561 AY620354 Marine sediment 7-2.1 Ϭ͘ϳϴͬϲϲͬϬ͘ϳϲ ϭͬϵϳͬϭ AY620320 Uncultured cercozoan A11 AY620322 Uncultured cercozoan D10 Ϭ͘ϴϵͬϳϭͬϵϵ AY620321 Marine sediment A14 Ϭ͘ϱϴͬͲͬϬ͘ϱϵ AY620281 0.5 FW F9 WƌŽƚĂƐƉĂŽďůŝƋƵĂ FJ824122 Ϭ͘ϳϯͬͲͬϬ͘ϳϯ DĂƚĂnjĂŚĂƐƟĨĞƌĂϱϱϴϵϱϲ AB275051 Uncultured eukaryote DGSM-51 AB275053 Uncultured eukaryote DGSM-53 Ϭ͘ϵϮͬϱϴͬϬ͘ϵϱ ďƌŝĂƚƌŝƉĂƌƟƚĂDQ303922 FN690390 Uncultured cercozoan 6c-F7 ŽƚƵůŝĨŽƌŵĂďĞŶƚŚŝĐĂ FJ824126 AY620356 Marine sediment 7-4.3 0.83/-/AB275052 Uncultured eukaryote DGSM-52 Ϭ͘ϵϲͬϱϰͬϬ͘ϳϳ AB275096 Uncultured eukaryote CYSGM-13 ‘eEbriida’ clade AB191410 Uncultured eukaryote TAGIRI-2 AB252750 Uncultured eukaryote NAMAKO-10 1 AY620348 Marine sediment 7-6.2 AY620314 Marine sediment 9-1.4 NC4 (Bass 2004) 0.99/98/1 JN090862 Uncultured eukaryote KRL01E2 WƐĞƵŽĚŽĚŝŋƵŐŝĂ cf. gracilis :ϰϭϴϳϵϰ Ϭ͘ϱϰͬͲͬϬ͘ϰϱ AB505500 Uncultured eukaryote RM1-SGM43 0.98/83/0.98 Thaumatomonas sp. (TMT002) DQ980486 M Euglyphida Marimonadida Clautriaviidae pt e d Nudifilidae Variglissida Discomonadida Ac Nudisarca 3 O D F R Q X G D , 0 % 5 , & $ 7 ( $ Ventricleftida ce Discomonadidae t Spongomonadida an Peregriniidae Esquamulidae Thaumatomonadida us cr ip Thaumatomonadidae 3 O D F R S H U O D Cryomonadida Matazida Ebriida Tectofilosida Y H Q W U L I L O R V D 7 + ( & 2 ) , / 2 6 ( $ Page 81 of 92 Figure1.2 0.76/72/0.52 AB534506 Uncultured eukaryote EF024287 Environmental Cercomonadida ‘eSarcomonad’ Clade Aurigamonas solis DQ199661 0.78/98/1 Agitata vibrans HQ121438 0.54/38/- Agitata tremulans AY748806 AY620288 0.5 FW F6 AY620287 0.5 FW F5 AY6203607-4.2 AY620289 6-3.7 AY620361 Marine sediment 7-3.4 AY620285 4-5.2 AY620290 FW leafy sediment PP1-14 AY620284 8-1.4 0.99/10/1 0.63/60/0.51/35/- 0.28/32/- 0.26/-/- NC7 (Bass & TCS 2004) x2 AY620273 Clade T glissomonad EU709264 Clade X glissomonad EU709259 Hetpan23 EU709256 EnvSoil Clade Y H32p118t7 EU709254 EnvSoil Clade Y H35p80t13 Sandona similis EU709173 0.85/88/0.80 0.99/99/1 Sandona limna HQ918177 Sandona ubiquita EU709153 0.63/33/0.83 Sandona mutans EU709138 Sandona heptamutans HQ918175 0.77/50/0.96 Mollimonas vickermani HQ918173 0.6/22/0.99 0.48/-/Mollimonas lachrima HQ918172 0.57/48/0.84 Flectomonas lenta EU709189 Flectomonas ekelundi AY496043 0.99/100/1 Neoheteromita hederae EU709211 Neoheteromita tolerans HQ121433 0.49/54/0.67 0.89/97/0.93 EU709217 Soil glissomonad Neoheteromita globosa U42447 0.53/-/0.99/100/1 Dujardina stenomorpha HQ918171 0.63/-/EU709276 Soil HetChi9 0.4/-/AJ130852 Clade U glissomonad 0.32/-/EU709273 Clade T glissomonad 0.99/100/1 Proleptomonas faecicola AF411275 0.87/-/ 0.78/-/Proleptomonas faecicola AJ305043 Bodomorpha prolixa HQ918170 0.99/100/1 0.31/-/Bodomorpha minima AF411276 EU709271 Clade Z glissomonad 0.98/87/1 EU709266 Het Aus17 EU709270 Clade Z glissomonad 0.2/-/EU647174 Environmental sediment 0.99/97/1 EU709250 Environmental soil Teretomonas rotunda EU709247 Allapsa mylnikovi EU709238 0.99/99/1 0.26/-/Allapsa ocior EU709230 0.8/37/0.97 Allapsa vibrans AF411265 ůůĂŶƟŽŶƉĂƌǀƵŵ CCAP1906/1 HQ918169 AB695472 Uncultured eukaryote 0.71/45/EF023937 Environmental glissomonad Amb 18S 1398 - Wrongly labelled in GenBank as ‘Eimeriidae’ EF023686 Clade Te glissomonad Amb 18S 94 EF024999 Clade Te glissomonad Elev 18S 4451 0.74/28/0.82 0.91/88/1 EF024516 ‘Thaumatomonadida’ environmental Elev 18S 823 Nucleocercomonas praelonga HM536165 0.99/86/1 B1c Paracercomonas bassi HM536163 Paracercomonas sp. FJ790736 Brabender et al 2012 Paracercomonas paralaciniagerens FJ790730 B1b Paracercomonas sp. Panama 101 FJ790734 Parcercomonas sp. Panama 107 FJ790735 (Brabender et al 0.48/85/0.99 Paracercomonas saepenatans FJ790731 0.7/41/0.88 2012) Paracercomonas sp. Panama67 FJ790732 Paracercomonas virgaria AY884340 Paracercomonas crassicauda FJ790725 Paracercomonas oxoniensis FJ790724 0.99/21/1 Paracercomonas compacta FJ790723 Parcercomonas minima FJ790720 B1a Paracercomonas producta FJ790721 (Brabender et al 0.94/81/0.94 Paracercomonas marina AF411270 0.46/41/0.66 0.99/-/0.56 Parcercomonas elongata FJ790729 2012) Paracercomonas astra FJ790727 Paracercomonas proboscata HM5361 0.56/-/Paracercomonas ambulans FJ790728 Paracercomonas Įlosa FJ790722 1/80/1 Metabolomonas insania HM536167 Metabolomonas sp. Panama83 FJ790737 Metabolomonas baikali HQ121440 1/99/1 DQ243992 Uncultured freshwater cercozoan PCB12AU2004 ƌĞǀŝŵĂƐƟŐŽŵŽŶĂƐĂŶĂĞƌŽďŝĐĂ AF411272 EF023473 Environmental cercomonad Amb 18S 1058 EF024294 Environmental cercomonad Elev 18S 695 Neocercomonas parincurva FJ790709 0.96/63/0.85 Neocercomonas jendrali HM536150 AB695470 Uncultured eukaryote MPE1-27 1/98/1 Neocercomonas giganƟca AY884320 FJ790752 Uncultured 13H.FBDBW Neocercomonas pigra FJ790707 Neocercomonas magna FJ790706 0.62/62/0.84 FJ790753 Uncultured 7A.IRDBW AY642694 Uncultured eukaryote p1.18 0.99/85/1 Neocercomonas 'plasmodialis' AF411268 Neocercomonas braziliensis FJ790702 0.99/88/1 Neocercomonas lata AY884325 Neocercomonas dactyloptera AY884334 0.99/82/1 FJ790705 Uncultured 13H.FDBW 0.99/84/1 Neocercomonas clavideferens FJ790704 Filomonas radiata FJ790712 FJ790756 Uncultured cercomonad 17-4.9.SA 0.99/59/0.98 Neocercomonas vacuolata FJ790711 Neocercomonas jutlandica AY496048 0.83/-/0.84 Neoercomonas celer FJ790710 FJ790754 Uncultured cercomonad 17-4.2.SA EF024850 Environmental cercomonad Elev 18S 1370 Cercomonas pellucida HM536148 Cercomonas directa HM536146 Cercomonas ambigua FJ790696 0.99/91/1 Cercomonas rapida FJ790694 Cercomonas deformans FJ790686 1/99/1 Cercomonas aī͘longicauda ATCC50317 U42449 Cercomonas mtoleri FJ790684 Cercomonas phylloplana FJ790689 Cercomonas lenta FJ790691 Cercomonas mutans FJ790692 Cercomonas laeva AY884321 0.99/85/0.99 Cercomonas diparavaria AF411266 EF023321 Environmental cercomonad Amb 18S 639 Cercomonas ellipƟca FJ790699 0.72/70/0.69 Cercomonas rotunda FJ790701 0.88/92/1 Cercomonas kolskia AY884330 Cavernomonas stercoris FJ790717 Cavernom mira FJ790718 0.78/97/0.99 AB534514 Uncultured eukaryote I 4 50 Eocercomonas sp. AF411269 (NCBI labelled ‘Cercomonas edax’) EF025030 Uncultured eukaryote Elev 18S 6781 Eocercomonas ramosa AY884327 0.99/98/1 Eocercomonas echina FJ790716 Eocercomonas minuscula FJ790714 0.99/100/1 Viridiraptoridae 0.99/100/1 Pseudosporida an us cr ip 0.91/54/1 Pansomonadida t 0.31/-/- Neocercomonas and Filomonas Ac ce pt e d M Paracercomonadidae 6 Glissomonadida $ 5 & 2 0 2 1 $ ' ( $ Cercomonadida Cercomonadidae Cercomonas Cavernomonas Eocercomonas 0.2 Page 82 of 92 Figure2 Ts z. zhukovi (T-3*) KF577807 Ts z. zhukovi (UPL1Be2) KF577810 Ts z. zhukovi (GMBGL1) KF577813 1/99 Ts z. zhukovi UPL1Bf2) KF577811 t 1/98 us cr ip Ts z. zhukovi (CCL4B) KF577812 0.88/67 Ts z. zhukovi (GMKL4) KF577809 0.87/73 Ts z. carolinensis (NC02) KF577814 zhukovi /seravini clade Ts z. paracarolinensis (NC03) KF577815 0.54/24 HM162157 Uncult ’Ascomycota’ clone T. zhukovi saxoni (HANTSF8) KF577808 0.47/16 Ts seravini seravini (T-2*) ATCC50636 Ts seravini varisquama (4e) KF577824 an Ts coloniensis (PML3A) KF577819 0.69/- Ts coloniensis (=vancouveri 3180W2) 0.99/98 0.95/84 EF138943 T. coloniensis (HFCC93) M Ts coloniensis (MLTB12) KF577820 0.98/90 coloniensis /constricta (lauterboni) /solis clade Ts oxoniensis (Hinksey) (CCAP1903/2*) 1/86 T. oxoniensis ( GMPL1) KF577821 0.81/48 1/95 d Ts oxoniensis (EP3) KF577822 0.7/- Ts constricta (CH3*) KF577826 1/99 Freshwater 0.84/- Ts constricta (BZ1) KF577825 pt e 0.92/62 Ts sp (P106) KF577827 Ts aff. solis (CH9) KF577823 HM240097 Uncultured ‘fungus’ clone ce 0.99/32 Env soil clade 1 Allas aff. diplophysa (ATCC50635) KF577828 0.61/- 0.95/66 Ac 5 I B V N B U P N P O B T Allas media (BKARH9*) KF577830 Allas Allas multipora (GSPB9*) KF577829 JQ666760 Uncultured soil ’fungus’ 1/97 DQ420715 Uncultured ’fungus’ Env soil clade 2 DQ420716 Uncultured ’fungus’ Reckertia gemma (VWB1thas*) KF577831 0.81/39 Reckertia filosa (AllasSN*) KF577832 Scutellomonas patella (CA05*) KF577835 Reckertia Scutellomonas Thaumatospina arabica (ThxDubai|*) KF577834 Thaumatospina mexicana (ThxMex*) KF577833 Thaumatospina 0.08 Page 83 of 92 Figure3 Output of sir_graph (©) mfold_util 4.6 Output of sir_graph (©) mfold_util 4.6 g a u g c g c a u c g u a a c u c g 20 c u u a a a g g a u a u a u a a a u 5’ c u a a a c a g c g c 260 u a u a c a c a g u u c c g c u a g u a u u u u c a c a c u u c u g a a g c g g a g u g u a c a u g u a g c a c u c g a g u c u c c u u a c c g a u u g u u c u g u u c g 40 c u a u g g c a u g u u g u a u g g g c u c c g g a a a a g c g a u u c g a u a u g u a g a u c c a a a u u c c g u u c a u g c u u a u c g c u c u u a a c c u g u a a g u 280 g c u u u u c g u c c c g a a g a u c u a c a c c g 260 g u u c g c g g u u u a 240 u a 240 a c u u c g g u c u u g g g a a u c g a u u a u u a g u u u c g u a 160 c u u u c c 120 g g g a g a g a a a a a c c u c u u u u g u a a c g a u c u u c u u u c c u a c a u a a c c 100 g u a a a u g u 140 u c c u u c g a a g u a Ts oxoniensis (Hinksey) a g c g c c g a c a g a u g a u a u a c a u a a u c u c a a g u c a c a u a c a g 220 c g a u a u a a g c 160 u a g c a u g c u u a a g c c c a c a u c g a u 200 u a g u g c u u a a a c g c u a c u a 180 c g g c c g u c u g u g 160 Enlarged loop a a u u u g a g c u u a c g c a a u c g a u u a g u g c u u a c a 200 u c a a III 180 g u a c u a c g g c c c u g u g u g c u dG = -91.13 [Initially -96.20] 4e a u g c u u g g a a c g c g u a u c u u u u g c a g c c u a c u c c a a a a u 220 g c a g a u a g a u u dG = -90.28 [Initially -92.80] T3_Denis a g a g c u c c g u u a u 160 u c g c a c u a Ts seravini seravini (T-2) a u u c g a g g M u g g a a u c 280 a g a a a a c a u a u a 120 a u u u c u g c g a g u a a g c c u a a 3’ c u g c a 5’ u c g u a u g u c a u 5’ c c g c u a c g g u a u 140 a u a a g a u g a u c a u a a u 220 g 3’ c a u g a c 20 140 u u g a u g u 80 u g a u g u a a u u 60 g u a u g a a g a g a u c c a g u a a g a u g u g u u 80 c u u c c dG = -88.31 [Initially -93.60] CH3 u g u u a g c u u u a g u g a a u u c u u u u c Created Tue Mar 5 09:01:23 2013 g c g u u u g g 240 Output of sir_graph (©) mfold_util 4.6 u a a g g Ts seravini varisquama (4E) u c a c c g u 120 a a g a u a a u a a u u g c c c a u c a 260 u c g u u a a c a u u c g g g u c g a g c u a u a a c u c u c c g u c IV 180 u a c c 100 g g g c g g a 120 g Ts zhukovi zhukovi (T-3) u c c a u 260 a g c a a g c g a u c u u g c c c u g u c g c g a u a u c g a u a g a u u a 240 a c g a u u u a u c c 100 g c g u a c u u u g g a u a a c c u a a g a a a c u u g a a a g 280 160 a c g c u a g a u u u g c g a a u g u u a a u g a c a u a c a c g g u c c c a g Created Mon Mar 11 05:08:00 2013 g a u c a g u c c c u Output of sir_graph (©) mfold_util 4.6 c a a 3’ u g c u u u u u g u u a a c c a u 200 40 a u a 5’ c u g g u g a g u a 140 u u g a c g a c u u c c g u g a a 5’ 3’ u a c u g u u u c g u a c u a a a a c a c a g a c c a g g u g a a u u g c u a a a g 80 u a 120 a 220 c g g u c g g c a u c c 80 u u g c u c u Ts constricta (CH3) c c u g c a u a u 20 u g u c g g c g g 40 g g c a a c a a a u a u c g u g a u a u a u c u g u a a u g g g a u c c u a a c c u c u u g a u g u a u u g u 100 u a 240 u g c a u a a c g u u c u u g a c u u c g c u u u g g c a a c c g u u c u g u g c u u a a u a c u u u u a g c g c g g u a u c c g u a 280 u g g u a c a u g c u 20 a u a u u a u a g g g c u c g c c 40 u a u a a u a 3’ a u g c g u a c g g c c a u c g 60 g a u a u a a u c a g g c a u 40 g c 80 g u c u u g c a I u g g u g 60 u u c g u u a g g u g c a c u g a u u a u c u g c u a II g u a u Created Tue Mar 5 09:23:05 2013 c u u u t u 60 c u c a an u Created Tue Mar 5 08:57:26 2013 u c u us cr ip c u c a c g a u c g a g a u c g c u g c u u a g c g u a a A u a u a c u c u a c 200 u a a a 180 g c c c u u a c g c g g c u u c c g c g Thaumatomonas g a a c c g 200 u u g u c u c g g u g c g c a a c c u c a u c a u g u 80 u g g g c a g g a c a a u a g g g u g c c a u g u a u c 20 c g u a c c c g g g g g u g a u c a g c c u a g a u 20 80 g a u g c c g 80 c 60 g c g a g u c g u g c u u g c g u u g g c 40 c g 20 g u u a g c g u a u u a u c a a c u u g g g g g u g g 40 c c u g a u g g c u c g 60 u u g a c c g g u c u u 40 g g u c c g a c u c g u g u g g 60 a a u u g g c g c c c g a u u pt e a c g u u u a a u a Output of sir_graph (©) mfold_util 4.6 u u c u Output of sir_graph (©) mfold_util 4.6 u u Output of sir_graph (©) mfold_util 4.6 c c a c g d dG = -96.28 [Initially -100.50] T2_Denis dG = -96.06 [Initially -100.60] T_oxoniensis a a a u g u a g u a g c a g u c g a u a c c a g c a a u u g c g a u g g 5’ u u a g a g a g u 5’ a u a a a c c g g g u c c a a a g c a c u u c g c c g u a 240 a g c g c a a u g u g c a 100 u u g u u c c c a u g c u u g u u a c c g g c g g 220 g g g c g c u u c a a c c u c c c g g g g g c g u g c u c c u 120 a u u u u g 120 g c u u c a g u u a u u c g a u g c a a g u a u a a u a g a u u g c u c a u c a a a c c c u c a u a 180 u u c a u a g g u g g a a c u u c Created Tue Mar 5 09:42:07 2013 180 c a c u g a a u u g Output of sir_graph (©) mfold_util 4.6 160 g g u c c c c u u u g c g u 160 u c g a a u a u Created Tue Mar 5 09:53:26 2013 a u a g C a u c Created Tue Mar 5 09:55:01 2013 a c u u c u u g c u u u g u c u u u g 200 g c u u a a u a a u a c g c g 240 g c Created Tue Mar 5 09:47:02 2013 200 c c c g g c g c gu c g c c g a a g u c c (ThxMex) a g g u c a c c c a g 160 c u u u c u g g u u g c u c g Thaumatospina mexica 140 c u a c c c 180 g u u g a u 140 c c g c g a u c g u c g c u u c u c g c c c u u u u c u a c a u 220 g c u a u c g g u u g u u a a u a c c c g u c g u a c a g u u c u a g c a g g 220 u u u u a c g u u c g u g u 140 a g u c u g c g g a c c u u g c g c u u 200 a u a c u u c u g g g u g u c c c c u u g u a g u c u u u a a c g g g c u u a u c c g g a u a u 120 a u g c c a g u u u g c a a c a a u 240 u a g u a c g c a 260 c a a c g c g u a c g u c 100 u a u a g u g c g ca c a a c g c 260 g u a c g 100 a c uu cc ac a auuuccca u g u c ce a c 3’ g a a a u Ac a u 5’ u u c uu aa ac a a auuucc g 280 c 260 (ThxDubai) u a 3’ g u 280 u c a u c a a c u a a g c a u a u u dG = -87.53 [Initially -92.80] ThxMex a g u 3’ Output of sir_graph (©) mfold_util 4.6 u dG = -84.53 [Initially -89.80] ThxDubai a a (VWB1Thas) B Thaumatospina arabica u dG = -87.60 [Initially -90.20] VWB1Thas ZĞĐŬĞƌƟĂŐĞŵŵĂ Created Tue Mar 5 09:28:16 2013 u c g c u a u a u g c g u a c g u a 60 g c u g u u c a a g u a u a u a u c u g u u g c c g a u a u g u u g u a g c g u g c u c g g 20 a g c a 80 u g 40 a a u Output of sir_graph (©) mfold_util 4.6 Created Tue Mar 5 09:31:04 2013 a a c c c g c a c g c c g g a a g a u u g u c g u a u g u 3’ u a u a a a g a a c c u g u u c u g g 100 u c u 260u u c g c u a a u u g u c a u u g u u g u u c c u g 280 u g a a a a c g u u g u u c c a u u c g c a g a u g u a g c a 260 a g a u g c c g u c g u u u a a c c 240 c u u g g u c c u c a g g a g ůůĂƐŵƵůƟƉŽƌĂ c a a 160 a u u g c u a u u g c c c u u c c g g g a a g g (GSPB9) c g a c c g a u c u g u u u c c c g a g c u u a g c a u c g c g u u g c u g c 180 u a 200 a a g u c u a c u a u a c g c g c g c g c c u u g c g c c u u g c g c u u a 120 c g u g u c c a c a u c g u u g u c c c c u u c c u a g a a g g g c a c c g c c g 200 g c u g u g g a 140 u c g g a c c g u g a a g u c c u g a a c c u u a u 180 a u a g c 200 a u u u a 100 c u c a c g g a u g g g g c g c c a a u a c g u (CA05) u a c a c u a u u g g u a Scutellomonas patella a g u c c g 220 u a a a a u a 160 u u a 240 c a g g u c u a c c cu a ugu a g u u a g 220 c g u u g u g 220 g u g g u a g u a u u u c g a u g c c a c g u u c c c a c 140 u c u c u c c g c g D u a c g u u a c a u a 140 u (BKARH9) g c u c u g a a a a c g 120 a g g g 260 g a a a a u 5’ a g a g c u u 240 u a g a g c u c a a u c c g g c g u c u u u 100 c a u u a Allas media c c a a c g c a a a a u g c a u a a a u 3’ a u c a g u u a c c g 120 u u c g c u u g a c u g a a a a a g u a u c a c u a u u a 5’ u 280 u 80 g a g c a a u g u g u a u a c a a u g c g g u c a g u a a u a c a a a c a c g u g u a a c c c g g a u u 5’ g c g a a a g c g c u g a u u g g c g u u 40 u 40 g c 80 u g g c g 20 c a c a g u g g a u u c c g u u g g g c u u c u a u g a a g u 20 60 u g a c a c u g a u u a u c u g g c u c g a a g u c g c 3’ c a u 60 a c a u a u a u g g u c a u a a c c u u u c E 180 u g a g c c a u a a c c u c g c 160 u c a g g c c u g u u g c dG = -94.38 [Initially -97.90] GSPB9 dG = -96.14 [Initially -103.00] BKARH9 Page 84 of 92 Ac ce pt ed M an us cr ip t Figure4 Page 85 of 92 Ac ce pt ed M an us cr ip t Figure5 Page 86 of 92 Ac ce pt ed M an us cr ip t Figure6 Page 87 of 92 Ac ce pt ed M an us cr ip t Figure7 Page 88 of 92 Ac ce pt ed M an us cr ip t Figure8 Page 89 of 92 Ac ce pt ed M an us cr ip t Figure9 Page 90 of 92 Ac ce pt ed M an us cr ip t Figure10 Page 91 of 92 Figure11 Thaumatomonas zhukovi/seravini clade Thaumatomonas coloniensis/lauterborni/solis clade lauterborni coloniensis and/or lateral perforations lost Allas lower tier only becomes oval multiple end struts single curved end struts Thaumatomonas oval or triangular lower tier, trampoline-like upper tier upper-tier lateral perforations; curved struts spine us cr ip t Thaumatomastix ? asymmetric spine scales ciliary scales lost bobbin-base spine scale igloolika filosa type perforated upper tier perforations lost Reckertia upper tier variations and splendida type flange spine and ciliary scales lost bobbin-base spine scales lost strut plate scale an and single-tier anterior ciliary scales Scutellomonas oval Thaumatospina M ciliary scales and Ovaloplaca ancestral triangular plate scale te d symmetric flanged, bobbin-base spine scales and end struts only Ac c mitochondrion ep Thaumatomonadidae upper tier added by mitochondrial shaping Peregrinia symmetric peripheral lattice ancestral two-tier scale Thaumatomonadida Gyromitus Peregriniidae scale rudiment Golgi-derived membrane ancestral single-tier scale Euglyphida Figure 11 Page 92 of 92