View other article » Take our survey »
Elsevier logo

Volume 173, Issues 1-4, September 2009, Pages 191-200

Precambrian Research journal image

Phase contrast synchrotron X-ray microtomography of Ediacaran (Doushantuo) metazoan microfossils: Phylogenetic diversity and evolutionary implications

  1. Corresponding author.
  2. a LPS of Nanjing Institute of Geology and Paleontology, Institute of Evo/Developmental Biology, and State Key Laboratory of Pharmaceutical Biotechnology, Nanjing University, Eastern Beijing Road, Nanjing 210093, China
  3. b Department of Earth Sciences, University of Southern California, Los Angeles, CA 90089, USA
  4. c Division of Biology, 156-29, California Institute of Technology, Pasadena, CA 91125, USA
  5. d Institute of High Energy Physics, Beijing 100049, China
  6. e Kewalo Marine Laboratory, University of Hawaii at Manoa, Honolulo, HI 96813, USA
  7. f European Synchrotron Radiation Facility, 6 rue Jules Horowitz BP 220, F-38043 Grenoble, France
  8. g Photon Factory, KEK, Oho 1-1, Tsukuba, Ibaraki 305-0801, Japan

Research highlights

  • Microfossils from the Ediacaran Weng’an Phosphate Member of the Doushantuo Formation have been studied using propagation phase contrast synchrotron X-ray microtomography.
  • Studies of Doushantuo embryos demonstrate the existence of a large suite of features found in modern embryos.
  • The diversity of the embryos suggests that the metazoan fauna of the Doushantuo included animals of poriferan, cnidarian, and both protostomial and deuterostomial affinity.
  • If this is correct, the last common ancestor of the bilaterian metazoan lineage, as well as that of sponges, cnidarians and bilaterians, pre-dated deposition of the Doushantuo strata.

Abstract

Microfossils from the Ediacaran Weng’an Phosphate Member of the Doushantuo Formation (Guizhou Province, southern China) have received widespread attention. The Doushantuo, which overlies the glacial deposits of the Nantuo Formation, was deposited following the Marinoan glaciation, the last extensive glaciation of Snowball Earth. Radiometric age dating indicates that the Doushantuo is older than 580 my, and hence that these microfossils are older than the Ediacara Biota. However, the diversity represented by these fossils has yet to be fully documented. A recent technological approach that has increasingly been used to image fossils, propagation phase contrast synchrotron X-ray microtomography, has allowed non-destructive study of both exterior and interior features of a variety of Doushantuo microfossils from the gray facies of the Weng’an Phosphate Member, cropping out along the axis of the Mt. Beidou anticline. Studies of Doushantuo embryos demonstrate the existence of a large suite of modern embryonic features, including macromeres and micromeres, cell lineage, polar lobes, compacted epithelia, equal and unequal cleavage, blastulation and gastrulation, and chorionic protection. Because embryos such as those here studied provide only a limited amount of phylogenetic information, and because adult metazoans of the types that produced these embryos have yet to be discovered in Doushantuo-age rocks, these fossilized embryonic forms can at present be assigned only to the various superclades represented amongst living Metazoa. The diversity of the embryos here studied suggests that the metazoan fauna of the Doushantuo may well have included animals of poriferan, cnidarian, and both protostomial (representatives possibly of basal protostome lineages) and deuterostomial affinity. If this interpretation is correct, it would then follow that the last common ancestor of the bilaterian metazoan lineage, as well as the last common ancestor of sponges, cnidarians and bilaterians, pre-dated deposition of the Doushantuo strata.


1. Introduction

Modern molecular phylogeny based on calibrated protein divergence rates provides predictions as to the depth in time of the Precambrian divergence of the remainder of the metazoa from calcisponges, of the cnidarians from Bilateria, and of the separation of the major bilaterian clades, the deuterostomes, ecdysozoans, and lophotrochozoans. Depending on mode of computation these divergence points fall either within the Ediacaran, i.e., following the Cryogenian Marinoan glaciation, which ended about 635 my (Hoffman et al., 2004; Condon et al., 2005) (Peterson and Butterfield, 2005), or 100–200 my earlier than the Marinoan (Douzery et al., 2004). Given this focus, paleontological and fossil biomarker evidence that might illuminate these fundamental early divergences is of tremendous interest (e.g., Brocks and Butterfield, 2009). Recent evidence on the occurrence of sponge biomarkers indicates an origin of sponges at least in the late Cryogenian, with a robust presence through the Ediacaran (Love et al., 2009). The earliest macroscopic paleontological remains, those of the Ediacara Biota (middle to late Ediacaran) have not much illuminated the basic divergences that gave rise to crown group metazoan clades; most species of the Ediacara Biota seem not to resemble any currently recognizable phyla (or their stem lineages) (e.g., Narbonne, 2004; 2005; Xiao and Laflamme, 2008), while others, such as the relatively late Kimberella, are already highly evolved animals, likely of molluscan affinity (Fedonkin and Waggoner, 1997). Occurring prior to the Ediacara Biota (see below for age discussion), the Doushantuo microfossil assemblage provides an essentially unique paleontological resource. Though there have been many studies, the interpretation of Doushantuo microfossils has posed difficult problems. Until very recently their morphology has been inferred either from petrographic thin sections, one per specimen, or from external SEM examination. Many of these microfossils are too simple in form to convey significant morphological information, as they resemble eggs or very early cleavage stages, although intriguing insights have been made on early development (e.g., Chen et al., 2006). Possible adult forms are rarely preserved, so far confined to microscopic sponges (Li et al., 1998), cnidarians (Chen et al., 2002; Xiao et al., 2000), and a single more complex bilaterian form, Vernanimalcula (Chen et al., 2004b; Petryshyn et al., 2009).

As often happens, the advent of a new technology can dramatically alter the state of knowledge, and such is the case with the application of Synchroton X-Ray Microtomography (SR-μCT) to fossil imaging (Chaimanee et al., 2003; Feist et al., 2005; Tafforeau et al., 2006; Donoghue et al., 2006). This technique is becoming more and more widely used in paleontology. The first works of this kind related to Weng’an microfossils were only very recently reported (Chen et al., 2006; Donoghue et al., 2006). Microtomography permits non-destructive computational examination of the specimen from any vantage point, visualization of internal characters in virtual sections in any plane, as well as 3D virtual extractions of internal structures. A third generation synchrotron can produce parallel, high flux, partially coherent monochromatized X-ray beams. Thanks to the partial coherence, it is possible to use propagation phase contrast based imaging techniques (PPC-SR-μCT), that can reveal many structures that are invisible, or hardly visible on classical absorption contrast based microtomography (Tafforeau et al., 2006; Friis et al., 2007; Lak et al., 2008). In the case of the Weng’an fossils it can reveal complex internal features of structures that might have been ignored or misinterpreted because of their deceptive exterior forms. Perhaps most important, it invites analysis of the small minority of Doushantuo embryos which are intrinsically most revealing: those representing later stages, with more complex structure, and larger numbers of cells.

Here we describe PPC-SR-μCT examinations of a new set of Doushantuo microfossils, chosen on two criteria. First, they are complex and distinctive enough to be potentially phylogenetically informative; second, almost all of the types discussed in the following exist in multiple copies in our analyzed collection, so that they can be treated as near-identical individuals of the same species. We regard this work as a small step towards the very large objective of developing an evidence-based estimation of the diversity and phylogenetic affinities of Doushantuo metazoans.

2. Past studies of Doushantuo fossils

The study of Doushantuo microfossil embryos now has about a 10-year modern history. Initially attention was focused mainly on algae and acritarchs, but more recently intense interest has arisen in Doushantuo animal embryos. The key, interrelated, issues are the diversity of animal forms represented in the microfossil assemblage as it continues to expand, and their possible phylogenetic affinities. Of great interest has been the report of a microscopic putative adult bilaterian, Vernanimalcula, which has been controversial (Chen et al., 2004b; c; Bengtson and Budd, 2004), and for which only a few specimens, from single sections, have been found (Chen et al., 2004b; Petryshyn et al., 2009). The current status of knowledge on other Doushantuo microfossils can be summarized as follows.

2.1. Cnidarians and sponges

There is significant evidence for microscopically scaled embryonic and adult stem group cnidarian forms. The most direct evidence refers to probable adult tabulate anthozoan forms which have been recorded in both thin section (Chen et al., 2002; Xiao et al., 2000) and in SEM analysis (Xiao et al., 2000). Chen et al. (2000) also reported a hydroid like form, and possible cross-sections of anthozoan polyps and stalks. Seventy-two specimens of similar gastrula-stage embryos that appear cnidarian, of tightly distributed size and morphometry were also reported by Chen et al. (2002), and the same exact forms were discovered earlier by Xiao et al. (2000). In addition, microscopic (<1 mm) putative postembryonic sponges were identified by the dense presence of acid-resistant monaxial spicules within an irregular shape filled with fossilized inclusions (Li et al., 1998).

2.2. Embryos of likely animal affinity

Chen et al. (2006) studied numerous microfossils resembling modern polar lobe-forming embryos. The specimens they studied were interpreted to represent a developmental sequence similar to that observed in lobe-forming embryos of modern spiralians (Chen et al., 2006). Xiao et al. (2007a) recently reported on a set of spirally decorated, large spheroid (∼500–890 μm) microfossils. These were interpreted as post-blastula stage embryos. However there is no internal structure and these forms could also be 1-cell eggs according to the authors.

Perhaps the most common of all fossil embryos of likely but heretofore undefined animal affinity is a solid blastula composed of approximately equal sized, polygonal cells, that is covered by a chorion-like structure and which we refer to below as “soccer balls” because of the stereotypic arrangement of their blastomeres at the 8- and 16-cell stages, which have an external pattern reminiscent of (although not identical to) a “soccer ball”. These were first seen in SEM images by Xiao et al. (1998), Xiao and Knoll (1999) and Xiao (2002); both in thin sections and SEM images by Chen (2004); and were analyzed morphometrically by Hagadorn et al. (2006) using synchrotron microtomography. In addition there have been reported from both SEM and thin section observations a large number of one, two and four cell embryos (Xiao et al., 1998; Xiao and Knoll, 1999; Xiao, 2002; Chen, 2004), but the form of these is too simple to convey much phylogenetic information, and they have also been interpreted as colonies of sulfur oxidizing giant bacteria (Bailey et al., 2007a; b; see also Xiao et al., 2007b).

3. Geological setting

The Doushantuo Formation (Fig. 1) is widely distributed in southern China (Fig. 2) as well as in areas of northwest China. It is underlain by a glacial deposit known as the Nantuo Formation, which was deposited during the Marinoan glaciation. The Doushantuo thus represents the first deposits after the last extensive glaciation of Snowball Earth. The Doushantuo Formation in Weng’an, Guizhou crops out along the axis of the Mt. Beidou anticline (Fig. 2). It is well exposed in Wusi quarry (Fig. 2), having a thickness of 49 m, and consists of five different lithological units (Fig. 1). The Weng’an Phosphate Member of the Doushantuo Formation, from which the fossils in this study were collected, varies laterally in thickness and lithology on a scale of meters to tens of meters (see natural profile in Fig. 1).

Fig. 1

Fig. 1: Natural and diagramatic profiles of the lithological sequence for the Ediacaran Doushantuo Formation in the Mt. Beiduoshan area, Weng’an, Guizhou. The natural profile was photographed from the western part of the Baishakan quarry and the thickness of the lithological units was measured from the Wusi quarry. WPM1 represents the black facies, and WPM2 represents the gray facies of the Weng’an Phosphate Member; all the embryos in this study were derived from the gray facies.

Fig. 2

Fig. 2: Index map of fossil localities showing locations along Mt. Beidou anticline.

In the Yangtze Gorges area, about 600 km northeast from Weng’an (Fig. 2), the Lower Dolomite Member of the Doushantuo, which is interpreted as a cap carbonate, contains an ash bed that has yielded a U–Pb age of 635.23 ± 0.57 Ma (Condon et al., 2005; Zhu et al., 2007). The top of the Doushantuo in the Yangtze Gorges area contains another ash bed that yields a U–Pb age of 551.07 ± 0.61 Ma (Condon et al., 2005). Within the Doushantuo Formation at Weng’an above the basal dolomite are a lower and an upper phosphorite interval that are separated by the Middle Dolomite Member, which is topped with a minor hiatus. The upper phosphorite interval, which contains the rich Weng’an microfossil fauna, has been interpreted to be younger than 580 Ma through carbon isotopic and sequence stratigraphic analysis (Condon et al., 2005). However, the upper phosphorite interval has also yielded a Pb–Pb age of 599 ± 4 Ma for the Weng’an Phosphate Member (Barfod et al., 2002), which is the Doushantuo interval where the specimens in this study occur. In addition, a Pb–Pb age of 576 ± 14 Ma for the Upper Phosphate Member (Chen et al., 2004a) has been obtained, providing additional evidence that the Doushantuo fauna indeed predates the Ediacara Biota and is thus the oldest record of animal fossils currently known. In summary the balance of evidence indicates that the Doushantuo microfossils are older than 580 my, and hence are the earliest fossil assemblage displaying animal affinities.

4. Taphonomy

Exceptional phosphate preservation was largely unknown until the 1970s when it attracted attention through the discovery of the Late Cambrian Orsten fauna (Müller and Walossek, 1985). This fauna yields very finely preserved miniature arthropods with anatomical details and soft integument (Waloszek, 2003). Xiao et al. (1998) and Li et al. (1998) first reported the discovery of fine-scale preservation of Precambrian phosphatized animal microfossils from the Doushantuo Formation.

The Weng’an fossil fauna is from two different lithological facies of the Weng’an Phosphate Member, the black phosphate and the gray phosphate, which differ from each other in nature and quality of preservation (Chen, 2004). The lower black phosphates are more organic-rich than the upper gray phosphates. The black unit contains fossil preservation with exquisite fine biological structures. Because of its lower carbonate content, acid isolation of embryos from the black unit is difficult. Embryos are readily obtained from the dolomitic gray unit through acid isolation, with common preservation of a variety of internal and external features. All of the microfossils reported on in this study are from the gray facies.

Differences in both quality and nature of preservation between the black and the gray phosphates are likely due to differences in sedimentary and diagenetic environments (Dornbos et al., 2006). The stratigraphic surface on which the black phosphate is deposited is interpreted as representing a hiatus (Chen, 2004; Dornbos et al., 2006). The black phosphate was deposited on top of this surface during transgression and it is a typical condensed, sediment-starved deposit (Chen, 2004; Dornbos et al., 2006). This unit is rich in organic material and includes cyanobacteria and algae, suggesting that these organisms were flourishing in the area. The black phosphate facies not only preserves the cellular structures of algae and embryos, but also yields exquisite fossils of putative metazoan larvae and adults (e.g., Chen et al., 2004b). The higher energy gray facies was deposited during a relative shallowing upward from the black to the gray facies (Dornbos et al., 2006). Here phosphatic intraclasts which contain microfossils are typically surrounded by a dolomite matrix. It is likely that many of these intraclasts were transported to the gray facies depositional environment from the black facies environment during high-energy episodes, such as storms (Dornbos et al., 2006). Primary phosphatic crusts that developed on the seafloor are also found in the gray facies (Dornbos et al., 2006).

During diagenesis different components of Doushantuo microfossils were variably resistant to decay (Xiao and Knoll, 1999). Doushantuo microfossils exhibit relatively little to significant decay before being phosphatized (Dornbos et al., 2005). Primary structures were preserved through phosphatization, and additional features were added later in the form of phosphatic encrustations (Xiao and Knoll, 1999).

5. Materials and methods

5.1. Provenance of specimens

The Doushantuo Formation has been intensively quarried for phosphates at the Wusi, Baishaikang and Nanbao quarries. All the specimens for this study came from the gray facies of the Weng’an Phosphate Member from these quarries (Fig. 1, Fig. 2).

5.2. Sample preparation

The samples were either treated under acetic acid or prepared in thin section. The samples for acid treatment used for PPC-SR-μCT were broken into pieces a few centimeters in greatest dimension and then submerged in a 10% acetic acid solution. The thin sections used in this study are ca. 50 μm thick, which is thicker than typical lithological sections. The thicker slice allows for three-dimensional visualization at different planes of focus. Thin section images were obtained with a stereomicroscope (NIKON ECLIPSE E600) under direct transmitted and polarized light at 100–1000× magnification.

5.3. Propagation phase contrast X-ray synchrotron microtomography

Several dozen acid residue specimens were selected for PPC-SR-μCT to be imaged on the beamline ID19 at the European Synchrotron Radiation Facility (Grenoble, France). We used a detector based on a CCD FreLoN camera linked to a revolver microscope optic. Depending on the sample sizes, we used isotropic voxel sizes of 0.28, 0.56 and 0.7 μm. We used a beam monochromatized at an energy of 20.5 keV using a multilayer monochromator. In order to obtain phase contrast effect, we used a sample-detector distance (propagation distance) between 15 and 50 mm depending on the pixel size, and 1500 projections on 180°. The software VGStudioMax 1.2 and 2.0 (Volume Graphics, Heidelberg, Germany) were utilized for 3D data processing, segmentation and analysis.

6. Results: biological diversity

Results primarily from PPC-SR-μCT reveal that a heretofore unappreciated diversity of animals was likely present by Doushantuo times.

6.1. A common chorionated multicellular embryonic form recovered in successive developmental stages

As noted above and often cited, a common form is the 8- and 16-cell stage embryos which externally resemble a “soccer ball”. These are large (600–800 μm diameter), spherical microfossils, e.g. like that shown in Fig. 3A. Here on the basis of their unusual internal organization as revealed by X-ray microtomography, we propose these forms as probable sponge embryos. Fig. 3B shows a virtual PPC-SR-μCT section through a 16-cell stage, 50% of the way through. A central cell is surrounded by large equal sized blastomeres. The specimen displays perfect radial symmetry, as shown by the section in Fig. 3C which is a 90° rotation with respect to Fig. 3B. The regularity of this form is illustrated in Fig. 3D and E which represents another specimen, one of many of identical structure; again a 90° rotation displays the same central blastomere surrounded with peripheral blastomeres extending to the surface. This unusual cleavage form is found in modern demosponges as shown in the SEM image in Fig. 3F (Leys and Ereskovsky, 2006). We and others have reported this same structure in thin sections, as illustrated here in the example shown in Fig. 3G. Note that the blastomeres are often preserved tightly compacted.

Fig. 3

Fig. 3: Ediacaran multicellular embryos with chorion-like structures (A–E and G–P) from gray facies of the Weng’an Phosphate Member of the Doushantuo Formation. (A–C) PPC-SR-μCT external 3D rendering (A) and virtual sections (B and C) of a specimen at 16-cell stage (specimen number 4D6), showing: “soccer ball” structure externally and a central blastomere surrounded with a subequally sized and perfectly radial arrangement of blastomeres (both of the sections are 50% of the way through the embryo; (C) with 90° rotation in respect to (B)); (D and E) PPC-SR-μCT virtual sections 50% of the way through another embryo at 16-cell stage (4B10), showing a central blastomere surrounded by a perfectly radial arrangement of blastomeres ((E) with 90° rotation in respect to (D)); (F) SEM image of a blastula in the modern demosponge Halisarca dujardini showing a central blastomere surrounded by a radial arrangement of blastomeres (reproduced from Fig. 8.A in Leys and Ereskovsky, 2006). (G) Steremicroscope image from a thin section cutting through the middle of an embryo, showing an interior blastomere surrounded with a perfectly radial arrangement of blastomeres; (H and I) PPC-SR-μCT external 3D rendering of a partly preserved embryo (H) and partially enlarged view (I) at possible 8-cell stage (M002), with a tangential slice through the chorion, showing the three-dimensional nature of the chorion cells, with typically a single depression in each chorion cell; (J and K) PPC-SR-μCT external 3D rendering (J) and section at 50% of the way through this 16-cell stage embryo (K), showing: an external chorionated covered with a single depression on each chorion cell (J), and an interior blastomere surrounded by a perfectly radial arrangement of blastomeres (K); (L–N) PPC-SR-μCT external (L and M) and internal 3D renderings 50% of the way through (N) an embryo at 8-cell stage (4E7), showing: “soccer ball” external structure (L), chorionated cover with large chorion cells (L and M) and the space between the embryo and chorionated shell (N), (M) is a 180° rotation of (L); (O and P) external (O) and internal views 50% of the way through (P) an embryo at 16-cell stage (4A3), showing: chorionated shell with large chorion cells, and an interior blastomere surrounded with perfectly radial arrangement of blastomeres. Scale bars represent 400 μm in A–E, G and H, and J–P, 40 μm in F, and 150 μm in I.

These embryos are commonly observed to be covered by a chorion-like structure. This chorion interpretation is based on a variety of observations. In many modern sponge embryos the chorion is composed of a single layer of flat cells (Leys and Ereskovsky, 2006). A virtual section of a fossil embryo with a chorion structure is shown in Fig. 3H and I. In the enlargement of Fig. 3I, the thin but three-dimensional form of the chorion structure and its cells can be distinguished. The cellular nature of the chorion in these microfossils is further supported by the extremely regular reticular pattern of boundaries that in their extreme regularity resemble cell boundaries. In addition note the single depression in each cell, probably representing the original site of a cilium or flagellum. This specimen is a 4- or 8-cell stage, possibly broken on the side not shown here. A later stage specimen displaying the identical type of chorion is shown in Fig. 3J and K. In this specimen the chorion is broken on one side so we can discern the “soccer ball” arrangement of the blastomeres within (Fig. 3J), and when computationally sectioned the same specimen reveals the tell-tale central blastomere (Fig. 3K).

It is interesting that there are several similar biological forms represented within this class of embryo, as indicated by different chorion morphologies. The 8-cell embryo in Fig. 3L–N has a chorion composed of much larger, probable flat cells than that of Fig. 3H–K. It too displays the external “soccer ball” arrangement of blastomeres as can be seen where the chorion is broken (Fig. 3L). Note that it would be easy to misinterpret the structure of the chorion for that of the embryo. Still a third form of chorion is shown in Fig. 3O and P, which however has the same internal structure. These three different chorion cell morphologies suggest that they may each represent a different species.

The embryos shown here represent several different developmental stages, up to the 16-cell stage. Developmental stages such as these evidently identical to those of modern metazoan embryos are of course inconsistent with alternative interpretations that have been proffered for similar forms (e.g., Bailey et al., 2007a; b). Furthermore, in our extensive library of thin sections are many examples of later developmental stages with larger numbers of compacted blastomeres that appear to represent more advanced late cleavage sponge eggs.

6.2. Complex embryos

Embryos at more advanced stages are relatively rare but here we present several examples, clearly representing diverse biological forms. In Fig. 4A is a highly organized, multicellular hollow embryo which is formed as a 1-cell-thick-layer ball of cells. The successive computational sections of the embryo in the same orientation from one end (Fig. 4A) display two deeper internal views (Fig. 4B and Fig. 4C), which expose a large cavity that we postulate is a blastocoel with several small cells seen at the floor of the blastocoel. We interpret these as cells which have delaminated from the blastocoel floor. This specimen strongly resembles a cnidarian (hydrozoan) gastrula. Cnidaria display many forms of gastrulation but the fossil embryo is of the same architecture as the modern form in Fig. 4D.

Fig. 4

Fig. 4: An Ediacaran embryo interpreted to be at gastrula stage (A–C) (020) from the gray facies of the Weng’an Phosphate Member of the Doushantuo Formation, showing: PPC-SR-μCT external 3D rendering of the embryo from one end (A), and successive virtual sections of two deeper internal views (B and C) of the embryo in the same orientation, displaying several small cells on the floor of the blastocoel; (D) a modern form of cnidarian gastrula showing small cells on gastrula floor (reproduced from Fig. 56E, Plate XII in Mergner, 1971). Scale bars all represent 150 μm.

Two different specimens are shown in Fig. 5, both of which we interpret as late cleavage stages of a probable bilaterian form, in which one end of the embryo consists of a micromere cap and the other of large endoderm-like blastomeres. External views of the micromere cap and the opposite “endoderm” side are shown in Fig. 5A and B. Note the approximately symmetrical arrangement of the two rows of possible endoderm cells in Fig. 5B. A computational section of this embryo displaying the large endodermal blastomeres is shown in Fig. 5C and in Fig. 5D is a different internal view which indicates the structure of this embryo: the micromeres form an ectodermal covering extending down over the endodermal cells. The embryo thus resembles a bilaterian stereoblastula undergoing epiboly. Another specimen of the same general import is shown in Fig. 5E and F. Here again one side of an advanced embryo consists of a micromere cap and the other of much larger prospective endodermal blastomeres.

Fig. 5

Fig. 5: Ediacaran embryos at late cleavage stage of a probable bilaterian form from the gray facies of the Weng’an Phosphate Member of the Doushantuo Formation. (A–D) PPC-SR-μCT external 3D renderings of the micromere cap (A) and the opposite “endoderm” side (B) of an embryo (010) with likely symmetrical arrangement of the two rows of possible endoderm cells, and two successive virtual sections (C and D) that display the large endodermal blastomeres (C) and micromeres (D); (E and F) PPC-SR-μCT external 3D rendering of an embryo (014), showing: a micromere cap (E) on one side and on the opposite side larger prospective endodermal blastomeres (F). Scale bars all represent 250 μm.

Fig. 6 shows a remarkable, late cleavage stage of which the dominant characteristic is a gigantic, polar macromere (Fig. 6A), while the other side of the embryo (Fig. 6C) consists of a cap of small blastomeres arranged somewhat symmetrically, that resemble micromeres of different sizes. Two computational sections are shown. In Fig. 6B we see the giant macromere sectioned in the same orientation as in Fig. 6A, surrounded with closely applied micromeres. Fig. 6D is a section through the micromere cap. The compacted intercell boundaries are clearly visible (Fig. 6D).

Fig. 6

Fig. 6: An Ediacaran embryo (019) at late cleavage stage characterized by a gigantic, polar macromere, from the gray facies of the Weng’an Phosphate Member of the Doushantuo Formation. (A–D) Showing: PPC-SR-μCT external 3D renderings of the gigantic polar macromere (A) on one side and a micromere cap on the opposite side (C), and two virtual slices of a giant macromere surrounded with micromeres (B) and micromeres (D), which are sliced in the same orientation through the endoderm pole and the micromere cap respectively. Scale bars all represent 300 μm.

Among living forms the complex embryos in Fig. 5, Fig. 6 most resemble those of protostomial bilaterians [e.g., Nassarius obsoleta (Clement, 1952)].

6.3. A duet cleavage form

Another type of early developmental form reproducibly recovered in this microfossil assemblage displays a “duet” based cleavage pattern. Whereas in modern lophotrochozoans spiral cleavage is typically based on a 4-macromere platform supporting an offset micromere cap, the acoels and their allies produce two primary macromeres, the first progeny of which are two successive, orthgonally arranged pairs of micromeres [“duet” cleavage; Boyer et al., 1996]. Here we show that 4-cell and 6-cell duet cleavage forms, although rare, were also present in the Doushantuo. Additional later stages are to be presented elsewhere.

Fig. 7A–D and E–H shows two different 6-cell duet embryos. Note the acellular chorion-like structure largely but not completely covering both of these specimens. Computational sections are shown at progressively deeper levels from the side of the embryo where the large twin macromeres are located. In Fig. 7A we see sections only through these macromeres; in Fig. 7B the section traverses the first pair of micromeres as well, which are characteristically located at right angles to the macromeres. The second cleavage division plane was evidently rotated 90° from the first. In Fig. 7D the plane of virtual section lies above the macromeres and reveals only the four micromeres. An exactly similar embryonic architecture is shown in the specimen of Fig. 7E–H, which is viewed from the same side as is the previous one. Two 4-cell stage specimens of the same or a similar form are shown in Fig. 7I–L and M–P. These lack surviving evidence of a chorion but this could easily have been lost due to taphonomic processes or sample preparation. The external view in Fig. 7I shows two macromeres and Fig. 7J shows the opposite side of this specimen and two micromeres. Virtual sections of the two macromeres and the macromeres plus micromeres (Fig. 7K and L) show the typical duet cleavage pattern of this specimen. The external view in Fig. 7M shows the micromeres from the outside, and the next image (Fig. 7N) shows the same embryo from the macromere side. In this case (as in many of these specimens) the macromeres are not of identical size. A virtual section through the macromeres (Fig. 7O) displays the remains of numerous yolk platelets; but in Fig. 7P we see that the platelet density is lower in the micromere duet.

Fig. 7

Fig. 7: Ediacaran duet cleavage embryos at 4- and 6-cell stages from the gray facies of the Weng’an Phosphate Member of the Doushantuo Formation. (A–D) PPC-SR-μCT virtual sections of an embryo (4D4) covered by a chorion from the macromere twin side to progressively deeper levels, showing successively: macromeres (A), the first pair of micromeres that was rotated at 90° from the macromeres in their cleavage division (B and C), and four micromeres only (D); (E–H) virtual sections through an embryo at 6-cell stage (4F7) from the macromere twin side (E) showing progressively deeper levels to four micromeres only (H); (I–L) a 4-cell embryo (4G4) showing: external view from macromere (I) and micromere (J) sides, and two virtual sections through the macromeres (K) as well as macromeres and micromeres (L); (M–P) a 4-cell embryo (4G3), with external view from micromere duet (M) and macromere duet (N), and two sections of macromeres (O) as well as macromeres and micromeres (P), both displaying extraordinarily well-preserved numerous yolk platelets. Scale bars all represent 400 μm.

The unique structure of this duet form marks it as a significant component of the Doushantuo microfossil embryo assemblage.

7. Discussion

7.1. The nature of the evidence

The initial issue is whether the specimens shown here really represent metazoan embryos. This question divides itself into two parts: could they be of abiological origin; and if biological, could they belong to non-metazoan groups. An abiological origin can be excluded for the major types of specimen shown here, for the simple reason that only genetically encoded developmental processes can generate identically replicate morphological biological forms within the species. This is not a new point, as we have utilized it earlier in pointing out morphometric regularities among very similar specimens examined both in thin section and by SEM (Chen et al., 2002; 2006). Nonetheless, when this point is applied to specimens of the identical structures of the 16-cell embryos of Fig. 3, for example, it becomes obvious that it is irrelevant to assume accidental origins. Furthermore, the exquisite state of preservation of some of the specimens can no longer be doubted. This extends to cellular and subcellular features (to be discussed in more detail elsewhere). Note for example, the repetitive fine structure of the patterned chorions of the embryos shown in Fig. 3H–J and the fossilized yolk platelets, like those seen in Hagadorn et al. (2006), in the duet macromeres in Fig. 7, and in particular the different distribution of these platelets in the micromere pair of Fig. 7P. Taphonomic processes certainly had an effect upon preservation of these microfossils, such as an apparent diagenetic enhancement of cell boundaries seen in many specimens, likely due to development of thin phosphatic crusts, as well as some distortion caused by early decay (e.g., Fig. 5). But, these taphonomic processes do not significantly mask the primary features of these specimens.

As for the general affinities of these embryos, they display none of the dense arrays of cuboidal cell walls recognized in many earlier studies of Doushantuo algae (e.g., Zhang et al., 1998); nor do they present any fungal characteristics such as linear budding chains or hyphae, and they certainly cannot be colonial bacterial, since the latter are not found in developmental cleavage stages. We conclude that we are here dealing with real evidence of metazoan embryos. No doubt the present literature on these embryos, this included, barely scratches the surface, and much more evidence will be required before a realistic appreciation of the biological complexity of Doushantuo metazoan life becomes possible; we have here merely an initial installment.

7.2. Evolutionary implications of the evidence

In that there is good prior evidence for adult sponge forms in the Ediacaran (Li et al., 1998; Love et al., 2009), our proposal that the very common “soccer ball” embryos are chorionated sponge eggs is essentially corroborative. Similarly, two kinds of possible adult cnidarian body plan have been described (see above for citations), and so the finding of a likely cnidarian gastrular form adds to the weight of evidence for the presence of this metazoan grade; note that other forms of possible hydrozoan gastrula were also reported earlier (Chen et al., 2002). We cannot yet appreciate the variety of organisms of either group, but there is no reason to assume it was small. Even if this were all we knew, its significance is not to be disregarded. Consider, for example, the complexity of the chorionic structures of Fig. 3, which had to be a maternal product. Though sponges are relatively distant from the bilaterians, it is increasingly clear that in terms of their genomic toolkit and their developmental use of this toolkit, the cnidarians are a close sister group to the bilaterians (Matus et al., 2006; Putnam et al., 2007). If cnidarians existed in the Doushantuo, then the basic principles of development of complex animal body plans had already evolved. We may add to this picture the probable presence of bilaterian forms—those implied by the organized embryos of Fig. 5, Fig. 6, Fig. 7, those interpreted by others as well as ourselves as being of possible bilaterian affinity, and the Vernanimalcula specimen(s). Clearly one conclusion that may be drawn is that by Doushantuo times large components of the suite of modern embryonic mechanisms had evolved. These included at the morphological level, macromeres and micromeres, cell lineage, polar lobes, compacted epithelia, equal and unequal cleavage, blastulation and gastrulation, and chorionic protection. And these are but external manifestations of the underlying world of genetically driven induction, spatial specification, and differentiation.

We take the view that until fossils of the adult metazoans which produced the embryos are recovered, the most that can perhaps be achieved is tentatively to relate the reproducibly recovered forms to the various superclades into which current Metazoa fall. There can be no presumption or suggestion of crown group identifications. Thus, the most far-reaching conclusions that might be drawn are that the metazoan fauna of the Doushantuo may well have included animals of protostomial affinity (possibly including basal protostome lineages), deuterostome affinity, and cnidarian and sponge affinity. If this is indeed the case, there is a striking further deduction to be made. This is that the last common ancestor of the bilaterians, and the last common ancestor of cnidarians and bilaterians, lived much earlier. It would then be necessary to seek paleontological evidence for the fundamental metazoan divergences in pre-Marinoan time, as suggested by the divergence dates of the Douzery et al. (2004) analysis.

Acknowledgements

This research was supported by Chinese Academy of Science Grant KZCX3-SW-14; National Basic Research Program of China (Grants 2007CB815800, 2006CB806400); National Science Foundation of China (Grants 40432006, 40772001); 111 Project and 985-2 Project of Nanjing University; NASA/Ames grant NAG2-1541; and the European Synchrotron Radiation Facility. CJY acknowledges grant support from Caltech through the Gordon and Betty Moore Distinguished Scholar Program. GL was supported by National Science Foundation of China (Grant 10675140) and FG by the Camilla Chandler Frost Fellowship at Caltech. We thank Marty Shankland (University of Texas, Austin) for valuable discussions and identification of yolk platelets. Bill Schopf, Shuhai Xiao and an anonymous reviewer provided valuable suggestions for improvement of an earlier version of this manuscript.

References