Technical Comments

Comment on “Fossilized Nuclei and Germination Structures Identify Ediacaran ‘Animal Embryos’ as Encysting Protists”

See allHide authors and affiliations

Science  09 Mar 2012:
Vol. 335, Issue 6073, pp. 1169
DOI: 10.1126/science.1218814


On the basis of putative nuclei and endospores, Huldtgren et al. (Reports, 23 December 2011, p. 1696) propose that embryo-like Doushantuo microfossils are nonmetazoan holozoans akin to mesomycetozoeans. However, both size and preservation preclude interpretation of internal structures as nuclei. Moreover, the authors may have conflated two different populations; some specimens display a pseudoparenchymatous organization incompatible with a mesomycetozoean comparison.

We welcome Huldtgren et al.’s (1) demonstration that Ediacaran microfossils commonly interpreted as animal embryos display characters that may be holozoan symplesiomorphies. This finding is fully consistent with, and perhaps a predictable consequence of, their earlier interpretation as stem group metazoans (2). Huldtgren et al.’s claim that these fossils are nonmetazoan holozoans analogous to mesomycetozoeans, however, rests on two flawed interpretations.

First, Huldtgren et al. interpret tomographically discrete nucleus-like structures in the embryo-like fossils as true nuclei, arguing that the cells divided via closed mitosis (nuclear membrane remaining intact during cytokinesis) rather than the open mitosis (nuclear membrane disintegrated during cytokinesis) observed in living metazoans. By itself, this interpretation would not falsify the stem-group animal hypothesis because open mitosis, a feature that evolved independently in multiple clades, could have evolved after the fossils diverged from the main line of metazoan evolution. Two observations, however, convince us that the structures in question cannot be preserved nuclei. First, petrographic and scanning electron microscopy studies show that the structures are preserved as late diagenetic, void-filling, botryoidal cements (1, 3). The late diagenetic origin of the structures means that they cannot be molds of nuclei, because phosphatic molding on a nuclear membrane template would necessarily have occurred during very early diagenesis. Internal molds of cells with crystallites nucleated on or adpressed against membranes or walls, for example, commonly incorporate nanocrystals (4) distinct from the void-filling, botryoidal microcrystals that replicate the nucleus-like structures (3). These botryoidal microcrystals are secondary overgrowths on a preexisting phosphatic substrate, similar to late diagenetic structures in other Doushantuo microfossils such as Vernanimalcula (5, 6). Thus, the behavior of the nuclear membrane during cytokinesis cannot be inferred from these late diagenetic structures. Moreover, the sheer size (up to 200 μm) of these microstructures precludes their interpretation as germline nuclei. Only certain ciliates are known to have macronuclei approaching the size of nucleus-like structures in Doushantuo fossils (7), but as Huldtgren et al. were not arguing for a ciliate interpretation, the large size of the fossil microstructures requires another explanation. The purported nuclei, which according to Huldtgren et al. [figure S6 in (1)] maintain a constant size through successive cell divisions, are volumetrically ~104 times larger than cells interpreted as reproductive spores formed by repeated cell divisions; these spores are supposed to (but physically could not) have hosted the putative nuclei with a full set of genetic material. Thus, either the nucleus interpretation or the ontogenetic connection must be incorrect.

The second problem with Huldtgren et al.’s interpretation involves the multicellular structures viewed as part of the life cycle of the embryo-like populations. Peanut-shaped individuals are interpreted as germination structures, based on comparisons to the extant mesomycetozoeans Ichthyophonus and Rhinosporidium seeberi. Ichthyophonus does indeed form peanut-like structures entirely filled with endospores, but at least the interior cells in structures illustrated by Huldtgren et al. [figure 3, E to J, in (1)] are not spores; they are vegetative cells with distinct cell walls that form a pseudoparenchymatous thallus that developed through apical cell division. Exterior cell clusters appear to be detached, but this is a taphonomic artifact, as movies kindly provided to us by Huldtgren et al. show that the purported endospores are attached to the thallus through cellular filaments and are indeed surrounded by faintly preserved thallus cells. The result is inconsistent with any life cycle known for Ichthyophonus and is distinct from the zonal maturation of spores in Rhinosporidium seeberi (8). Moreover, the proposed ontogenetic connection between Megaclonophycus-stage fossils (~103 cells) and peanut-shaped fossils (~106 cells) leaves a large developmental gap that cannot be bridged by a common envelope because no such feature is preserved in the anatomical illustrations provided. Doushantuo fossil assemblages do include distinctly lobate, cellularly differentiated, pseudoparenchymatous thalli, but these have been interpreted as red algae (9). Thus, we believe that Huldtgren et al. have conflated two distinct organisms in their interpretation.

Stripped of the two invalid attributions, the Doushantuo embryo-like fossils can be accommodated by the stem-group metazoan interpretation but are inconsistent with a mesomycetozoean analog. Their ornamented envelopes can be compared with those of modern animal eggs (10) but are entirely different from sporangium walls of modern mesomycetozoeans (11, 12). Also, their cells form an obligate multicellular organization with stable and reproducible Y-shaped cell junctions (Fig. 1), distinct from facultative cell aggregation in clonal protists. Considering that partial degradation of modern embryos (13) and Parapandorina fossils results in cell volume reduction, rounding, and separation (Fig. 1E–F), it is remarkable that the Y-shaped junction is maintained even in shrunken cells (Fig. 1E). This indicates that this configuration is biologically stable and not an artifact of the physical confinement of cells within a rigid envelope. The biological formation of Y-shaped cell junctions requires flexible cell membranes and cell-cell adhesion (14), an important step toward complex multicellularity (15). Third, cell cleavage in Parapandorina (as in modern animals) follows a 2n sequence, whereas cytokinesis in most protists, including mesomycetozoeans, does not follow a 2n sequence because it commonly involves multinucleate stages (1618).

Fig. 1

Cell cleavage and taphonomic degradation of Parapandorina-stage fossils. (A to C) The formation of a Y-shaped cell junction requires cell membrane flexibility and cell-cell adhesion. The sketch illustrates the second round of cell division (14). (D) Well-preserved specimen with nearly intact cell size and shape. Note the Y-shaped cell junctions. (E) Degraded and shrunken cells within a partially preserved envelope. Note the Y-shaped cell junctions. (F) Degraded and rounded cells. The cell in front became spherical and began separation from other cells, whereas the cell in back still maintained a polyhedral shape. Scale bars, 100 μm.

We agree with Huldtgren et al. that some of the features used in support of the animal embryo interpretation (10) may be holozoan symplesiomorphies and that the Doushantuo fossils do not show all the features that collectively define crown-group Metazoa. However, we would expect a stem-group animal to have holozoan symplesiomorphies, lack some or most crown-group synapomorphies (19), and evolve autapomorphies and homoplasies. Indeed, it has been proposed that early animal evolution may have involved a synzoospore stage without embryogenesis (20). Given the problems regarding the nuclear and life cycle arguments advanced by Huldtgren et al., we believe that the obligate, stable, and reproducible multicellular organization of the Megasphaera-Parapandorina-Megaclonophycus complex places them on the animal branch of the holozoan tree, because animals are the only holozoans that have evolved this degree of multicellularity. Another possibility is that at least some of the microfossils attributed to embryo-like populations represent the distinctive reproductive propagules produced by the early branching animals such as Trichoplax (15, 21). This would be consistent with recent interpretations of some Ediacaran macrofossils as little more than upper and lower epithelia that lined a fluid-filled or mesogloea-like interior and fed by absorption or phagocytosis (15, 22, 23). As a final note, we draw attention to similar microfossils (Fig. 2) found in association with undisputed Cambrian animal embryos (24, 25). The presence of similar forms in association with Cambrian (and living) animal embryos does not prove that the Doushantuo microfossils are animal, but it does offer another avenue to further test the mesomycetozoean comparison.

Fig. 2

Comparison between (A) modern animal cysts, (B) Doushantuo microfossils, and (C to H) Cambrian fossils in association with the animal embryo fossils Olivooides, Pseudooides, or Markuelia from South China. (A to C) Similar ornamentation on modern animal cyst (A), Doushantuo fossil Tianzhushania ornata (B), and Cambrian microfossil (C). (D and E) Forms similar to lobate Doushantuo fossils. (F) Two spiny acritarchs (similar to the Doushantuo form Cavaspina acuminata) on an Olivooides animal embryo. (G) Enlarged view of (F) to show detail of processes. (H) Form similar to the Doushantuo form Asterocapsoides sinensis. Scale bars, 100 μm. [(A) from Belk et al. (26); (B) from Xiao and Knoll (10); (C) from Dong (25); (D) to (H) courtesy of X.-P. Dong]


Stay Connected to Science

Navigate This Article