Diplophrys, Barker, 1868

Anderson, O. Roger & Cavalier-Smith, Thomas, 2012, Ultrastructure of Diplophrys parva, a New Small Freshwater Species, and a Revised Analysis of Labyrinthulea (Heterokonta), Acta Protozoologica 51 (4), pp. 291-304 : 300-301

Contrast between Diplophrys and Amphifila marina

Conservation of the name Diplophrys for D. parva rather than A. marina merits discussion. Until an authentic culture of D. archeri is sequenced, we cannot be sure that retention of the generic name Diplophrys for the D. parva rather than the Amphifila clade is correct, but a decision one way or the other had to be made. We picked D. parva for three reasons: first because its more rounded, less pointed shape, is more like D. archeri than is the spindle-shaped A. marina . The consistent phylogenetic contrast between the elongated Oblongichytrium and round Thraustochytriidae sensu stricto ( Yokoyama and Honda 2007, and Fig. 6 ) shows

Ultrastructure of Diplophrys 301

that small differences in cell shape can have surpris- as occurs among some slime molds, but as its vegetaingly deep phylogenetic significance in Labryrinthulea. tive cell structure is dissimilar from slime molds, and Second are the filopodia: in D. parva and archeri they cell aggregation is well known as a polyphyletic charare branched but non-anastomising, and both show only acter, it should not be placed in Mycetozoa. Though its minimal cell motility if any – no locomotion was men- ultrastructure remains unpublished, Dykstra and Olive tioned in the original descriptions of D. archeri (Barker (1975) stated that it lacks sagenetogens and has thin 1868). By contrast Amphifila locomotes by active glid- scales. Sorodiplophrys vegetative cells crawl using ing and shows fine filopodial anastomoses, both char- contractile non-granular filopodia, whose contractilacters shared with Sorodiplophrys , but not D. archeri . ity makes them perhaps more similar to those in the Thirdly, D. archeri and D. parva are both from fresh- cercozoan superclass Ventrifilosa (Cavalier-Smith and water, whereas Amphifila is marine, and conservatism Karpov 2012), comprising the filose amoeboid classes of freshwater versus marine habitat is pronounced in Imbricatea and Thecofilosea ( Howe et al. 2011 a), than many protists ( Cavalier-Smith and Chao 2012), and also to Diplophrys . As Imbricatea often also have scales, it shows a non-random distribution across Labyrinthulea. is possible that Sorodiplophrys belongs in that class, One can argue that Labryrinthulea were probably an- which includes a variety of amoebae and flagellates cestrally marine. However, most lineages of the clade with similar contractile, non-granular branching filoto which Amphifila belongs are freshwater (or soil, eco- podia ( Howe et al. 2011 a, Cavalier-Smith and Chao logically cognate), so that clade was probably freshwa- 2012). Moreover, the testate Amphitremidae , with bipoter for most of its evolutionary history, and the ances- lar filopodia analogous to, but more robust than, those tor (or ancestors) must have made one relatively recent of Diplophrys , is currently assigned to Thecofilosea switch into the oceans, perhaps accompanying the sea ( Cavalier-Smith and Chao 2012). However, we have grasses with which it is commensal. The sequence clos- adopted the more conservative stance of retaining Soroest to D. parva comes from European peat bogs and diplophrys within Labyrinthulea, for two reasons. First, D. archeri was from Irish moors, both consistent with Dykstra and Porter (1984) noted thin scales of Sorodipthe morphological evidence that D. parva and archeri lophrys resembling those of Labyrinthulea. If the scales are mutually closer than to Amphifila . had been more like any of the diverse siliceous scales Scale ultrastructure, often good phylogenetic in- of the scaly taxa now placed in Imbricatea, they would dicators ( Cavalier-Smith and Chao 2012, Howe et al. probably have mentioned that and even more strongly 2011a), strongly supports this; we found that D. parva doubted its affinity with Labyrinthulea. Thus Sorodiphas oval to elongated scales (~ 1 µm) decidedly differ- lophrys is probably not an imbricate. Secondly, they ent in size and shape from the round scales (~ 2 µm) stressed that Sorodiplophrys is osmotrophic and not of A. marina . A second ultrastructural difference is that phagotrophic, also making it unlikely that it is a scaly D. parva has an obvious dense structure somewhat imbricate cercozoan amoeba (all have siliceous scales, resembling a sagenetosome, whereas no evidence for not unmineralized organic ones like Labyrinthulea). a sagenetosome was seen in Amphifila , in which respect In contrast to Amphifila and Sorodiplophrys , both also it resembles Sorodiplophrys (Dykstra and Olive currently recognized species of Diplophrys present di- 1975). agnostic features of the genus, i.e. ellipsoidal to ovoid cells, non-aggregating cells, enclosed by a thin enve- Increased diversity of Diplophrys -like protists lope (shown to be imbricated scales by fine structure Only two previously described species were re- analysis) with pseudopodia emerging typically from cently accepted as Diplophrys : Diplophrys archeri and two poles of the cell, forming a branching rhizopodial Diplophrys marina (here moved to Amphifila . Diploph- fan toward the periphery; there is at least one intracyrys stercorea described by Cienkowski (1876) was re- toplasmic refractive granule, presumed to be lipid. In assigned to a separate genus Sorodiplophrys (Dykstra D. archeri , the refractive granules (one or more) are and Olive 1975), with stercorea the type species. It is typically very prominent, yellowish in color, and oc- a sorocarp-producing protist, thus sharply distinct from cupy a large portion of the cell volume when viewed Diplophrys and Amphifila , despite having sufficiently by light microscopy. Published images of D. archeri similar vegetative cells to A. marina (net-like filopodia) are typically in the range of 10–15 µm or somewhat to make a relationship plausible. The sorocarp (stalk- larger (e.g. Barker 1868, Kudo 1977, p. 568). D. marina borne fruiting body) is a product of cellular aggregation cells (3.7–5.9 × 5.1–8.5 µm) are ovoid with round Gol- 302 O. R. Anderson and T. Cavalier-Smith gi-derived scales (1.5–1.9 µm). The Diplophrys -like phenotype had only three named species ( D. archeri , D. marina , and Sorodiplophrys stercorea ) prior to this publication. Their placement now in three separate genera and families better reflects their evolutionary diversity and should stimulate further research on this unique protist type – neither a rhizopod nor a fungus but a very distinctive, albeit neglected, osmotrophic phenotype. Many understudied protists are not in culture (e.g. D. archeri ), impeding molecular genetic analyses, but many more could probably be cultured with even a modest effort. Recent light micrographs of D. archeri with accurate diagnostic size and morphology for this species as described by Barker (1868) (e.g. http://starcentral. mbl.edu/microscope/portal.php?pagetitle=assetfactshe et&imageid=9704) show that D. archeri , and no doubt many genetically distinct look-alikes, can be isolated from the natural environment. Without a targeted study of Diplophrys , currently with only two verifiable species ( archeri and parva ), it is premature to judge wheth- er its taxonomic diversity is really limited to the three sequences that branch robustly together in Fig. 6 , or is much more extensive. However, given the small size of D. parva , and its broad similarity to the genetically very distant Amphifila , it is likely that many additional cryptic species will be discovered. The clade containing Amphifila is currently more speciose. Possibly one of the two distinctly deep-branching soil lineages in that clade is related to the dung-dwelling Sorodiplophrys , as dung dwellers are most likely to have evolved from soil biota; if that could be confirmed, it would make that quite speciose clade equivalent to the new superfamily Amphifiloidea . More intensive research on this microscopically distinctive but remarkably conservative morphotype is warranted.

Figures

Fig. 6. Maximum likelihood tree for 327 heterokonts emphasizing Labyrinthulea and Eogyrea of phylum Bigyra and the non-heterokont outgroups. Internal branches for the five clades at the top of the tree are collapsed, but are being published separately (Cavalier-Smith and Scoble in press); the numbers to their right indicate how many sequences were included in each. Bootstrap supports for bipartitions are based on 1,000 resamplings using GTRMIX option of RAxML. Black bullets indicate 100% bootstrap support. The sequence attributed to ‘Labyrinthuloides haliotidis’ might be from a thraustochytrid contaminant rather than from Aplanochytrium (=Labyrinthuloides) haliotidis (Leander and Porter 2001).