Comparative topology and structure of Thiomargarita and Parapandorina cell
clusters: Both Doushantuo and Thiomargarita diads exhibit an undisturbed division
plane (Fig. 1b, b’). Three-cell clusters, thought to result from the incomplete division of a
two-cell stage are present in both Parapandorina and Thiomargarita (Fig. 1c, c’).
Whether or not the undivided larger cell of the Thiomargarita triads later undergoes
reductive division to produce a tetrad is presently unknown. Both Thiomargarita and
Parapandorina tetrads exhibit a variety of cell configurations. Four-radiate cross-
junctions of division planes in tetragonal Doushantuo tetrads are shown in Figure 1d’, a
configuration common in extant non-mammalian embryos and in Thiomargarita tetrads
(Fig. 1d). Some Doushantuo octads are also observed to exhibit radiate cross-junctions
(Fig. 1c in 1), which undoubtedly resulted from division of tetragonal tetrads. Rhomboidal
tetrads, which are characterized by two opposite bodies in contact, and two opposite
bodies separated by a gap, are also observed in Parapandorina (Fig. 7.5 from 2,
reproduced in Figure 1e’), and in Thiomargarita tetrads (Fig. 1e). Perhaps a more
common topology observed in Doushantuo tetrads are decussate or tetrahedral
configurations, which result from deformation of preexisting cell-division planes3. Such
deformation is thought to require non-rigid cell walls/membranes and produces Y-shaped
triple junctions of division planes (Supplementary Fig. 1). Although bacterial cell walls
were once thought to be rigid structures, they are now known to be quite elastic4.
Deformation of division planes resulting in Y-shaped triple junctions are observed in
multi-planar Thiomargarita tetrads, including partially-deformed tetragonal tetrads
(Supplementary Fig. 2a), decussate tetrads (Supplementary Fig. 2), and tetrahedral tetrads
SUPPLEMENTARY INFORMATION
doi: 10.1038/nature05457
1www.nature.com/nature
(Supplementary Fig. 3). Y-shaped triple junctions are also observed in 8-cell
Thiomargarita and multi-cell Parapandorina specimens (Supplementary Figure 4, 5c,
5d). Although Thiomargarita from off the coast of Namibia are known to form chains5,
Thiomargarita from the Gulf of Mexico, the focus of this study, have not been observed
to form chains.
Thiomargarita cell clusters have yet to be observed containing hundreds of
internal bodies, such as Megaclonophycus, another globular microfossil from the
Doushantuo Formation. However, these microfossils have never been demonstrated to be
part of a developmental continuum with Parapandorina2. Megaclonophycus are loosely-
packed clusters that contain rounded, rather than polyganol, internal bodies2, and do not
exhibit a blastocoel, as would be expected in embryos with similar numbers of cells6, thus
calling into question their metazoan affinities.
Internal bodies: Some Parapandorina specimens include smaller subcellular
structures of non-diagnostic shape6. Such bodies are consistent with diagenetically-
altered inclusions of the type that are common in Thiomargarita, but as with other
hypotheses, their origin is ultimately ambiguous. A few Parapandorina specimens
(n=10) include larger spherical-to-reniform internal structures6. Hagadorn et al.
6 suggest
that these bodies might be organelles, however they also allow for the possibility that the
larger internal bodies resulted from inorganic mineral precipitation or shrunken
cytoplasm as observed in other microfossils e.g.,7
, including algae from the Doushantuo
Formation (Fig. 3H in 6). As a sheathed organism with vacuole contents that are
chemically-distinct from surrounding waters, Thiomargarita cells could also have
undergone cytoplasmic shrinkage or internal diagenetic mineral precipitation resulting in
2www.nature.com/nature
doi: 10.1038/nature05457 SUPPLEMENTARY INFORMATION
intracellular structures. Inclusions in Thiomargarita also occasionally form aggregates
(Supplementary Fig. 6), likely as a result of cytoplasmic degradation, that are similar in
size and shape to the large intracellular structures observed in a small number of
Parapandorina specimens. These aggregates often exhibit approximate symmetry across
division planes (Supplementary Fig. 6b), and can sometimes be found as pairs in each
cell of a multi-cell cluster, as is also observed in at least one Doushantuo tetrad6, though
these aggregates likely result from degradational, rather than physiological, processes.
Supplementary Figure 1: Thiomargarita tetrad exhibiting deformation of the
preceding division planes with Y-shaped triple junctions; (see arrows) as observed in
a cartoon modified after 3 showing an intermediate stage in the deformation thought to
result in many Doushantuo Parapandorina tetrads. Scale bar = 100 μm.
2www.nature.com/nature
doi: 10.1038/nature05457 SUPPLEMENTARY INFORMATION
Supplementary Figure 2: Thiomargarita tetrad in an approximately decussate
geometric configuration, with two cell pairs approximately at right angles to one
another, similar to the geometry observed in some Parapandorina tetrads (cartoon
modified after 3). The dark material surrounding the cluster is composed primarily of
small filamentous and spherical cells, with sizes and morphologies similar to the
filamentous and spherical structures commonly found on surfaces of Doushantuo
microfossils. Scale bar = 100 μm.
Supplementary Figure 3: Thiomargarita tetrad exhibiting an approximately
tetrahedral geometry akin to another geometry commonly observed in microfossil
tetrads from the Doushantuo Formation. Scale bar = 100 μm.
2www.nature.com/nature
doi: 10.1038/nature05457 SUPPLEMENTARY INFORMATION
Supplementary Figure 4: Thiomargarita octads exhibit Y-shaped junctions (white
arrows), similar to those observed in multi-cell Parapandorina clusters (after 2, Fig.
8.8). Scale bar =100 μm.
Supplementary Figure 5: Thiomargarita cell clusters (a: 3-cell cluster, b: 4-cell
cluster, c, d: probable 8-cell clusters) show a variety of geometries that indicate
deformation of the previous division plane, as observed in Parapandorina clusters.
All scale bars = 100 μm.
Supplementary Figure 6: Aggregates of internal inclusions, presumably resulting
from cytoplasmic degradation, appear as large spheroidal to reniform intracellular
bodies in a small number of a) solitary Thiomargarita cells, and b,c) multi-cell
clusters. All scale bars = 70 μm.
Comparative morphology of Tianzhushiana, Megasphaera inornata, and
Thiomargarita: The Doushantuo Formation contains abundant large spherical
2www.nature.com/nature
doi: 10.1038/nature05457 SUPPLEMENTARY INFORMATION
microfossils that exhibit a range of morphological features. Abundant indistinct smooth
spheres of uncertain affinities, generally categorized as sphaeromorphic acritarchs, are
often thought to represent algal resting cysts. Some Doushantuo spherical bodies are
encased in an envelope (Megasphaera8). It has been recognized that some of these
envelopes possess external surface ornamentation (Megasphaera ornata), while others
lack ornamentation (Megasphaera inornata)2. This ornamentation has been central to the
argument for a metazoan egg interpretation of Megasphaera2,9. Recently, it was proposed
that Megasphaera results from taphonomic alteration of the acanthomorphic acritarch
Tianzhushiana tuberifera10, which is characterized by external processes that penetrate a
multi-lamellate outer wall11,12. Similarities between the middle wall of T. tuberifera and
the outer wall of M. ornata, and the finding of an additional outer wall on a few M.
ornata specimens lead to the hypothesis that they represent the same species. Based on
the hypothesis that unornamented globular fossils may have once had an ornamented
outer wall that was not preserved, the unornamented M. inornata was also proposed to be
synonymous with Tianzhushania10. Given the great abundance of unornamented globular
microfossils in the Doushantuo phosphorites, we find this explanation insufficient to
explain most unornamented specimens. Therefore, we retain the use of M. inornata,
which we interpret as possible fossilized giant sulfur bacteria based on its abundance in
phosphorites. Deflated envelopes lacking internal bodies are also observed in the
Doushantuo microbiota2 (Supplementary Fig. 7a'), and similar deflation is commonly
exhibited in Thiomargarita cells with ruptured internal vacuoles (Supplementary Fig. 7a).
However, modern animal eggs and phosphatized Doushantuo acritarchs are also observed
2www.nature.com/nature
doi: 10.1038/nature05457 SUPPLEMENTARY INFORMATION
to exhibit deflation and collapse2. Ultimately, unornamented globular fossils are
taxonomically ambiguous and could represent more than one organism.
Supplementary Figure 7: Deflated Megasphaera (a’) (after 2) may indicate an
initially hollow interior, such as the large vacuole in Thiomargarita, which is
occasionally observed to rupture resulting in a deflated cell (a). Scale bar = 100 μm.
Abundance of Doushantuo microfossils: The unusual abundance of globular
microfossils in the Doushantuo Formation has long been considered problematic for the
animal embryo interpretation13; concentration via sedimentary processes has been
proposed as a possible solution14. Such a circumstance does fall within the realm of
possibility. However, the globular microfossils generally occur as grains within larger
reworked clasts 2,15,16. Literature on this topic suggests that the fossils were phosphatized,
size sorted by currents, cemented together, ripped-up, and re-deposited elsewhere within
the larger clasts15,16. Each stage of cementation followed by reworking and re-deposition
loses the memory of the former clast size. Such multiple reworking would more likely
dilute, not concentrate, the microfossils in the deposit. Thus, we do not find concentration
by sedimentary processes to be the most likely solution.
The interpretation of the microfossils as bacteria more easily explains their
abundance, as Thiomargarita is known to occur in great abundance (up to 200 g m-2),
2www.nature.com/nature
doi: 10.1038/nature05457 SUPPLEMENTARY INFORMATION
which would translate to approximately 107 cells per m-2. In the Gulf of Mexico,
abundance is somewhat lower at ~ 105 cells per m-2 (see Supplementary Fig. 8).
Supplementary Figure 8: Hydrocarbon-seep sediments covered with very abundant
Thiomargarita cells. The width of the pH microelectrode tip at center is ~500 μM.
1 Dornbos, S. Q. et al., Precambrian animal life: Taphonomy of phosphatizedmetazoan embryos from southwest China. Lethaia 38, 101-109 (2005).
2 Xiao, S. and Knoll, A. H., Phosphatized animal embryos from the NeoproterozoicDoushantuo Formation at Weng'an, Guizhou, South China. J. Paleontol. 74, 767-788 (2000).
3 Xiao, S., Mitotic topologies and mechanics of Neoproterozoic algae and animalembryos. Paleobiology 28, 244-250 (2002).
4 Doyle, R. J. and Marquis, R. E., Elastic, flexible peptidoglycan and bacterial cellwall properties. Trends Microb. 2, 57-60 (1994).
5 Schulz, H. N. et al., Dense populations of a giant sulfur bacterium in Namibianshelf sediments. Science 284, 493-495 (1999).
6 Hagadorn, J. W. et al., Cellular and subcellular structure of Neoproterozoicanimal embryos. Science 314, 291-294 (2006).
7 Knoll, A. H. and Barghoorn, E. S., Precambrian Eukaryotic Organisms: AReassessment of the Evidence. Science 190, 52-54 (1975).
8 Chen, M. and Liu, K., The geological significance of newly discoveredmicrofossils from the upper Sinian (Doushantuo age) phosphorites. Scientia Geol.Sinica 4, 317-324 (1986).
9 Xiao, S., Zhang, Y., and Knoll, A. H., Three-dimensional preservation of algaeand animal embryos in a Neoproterozoic phosphorite. Nature 391, 553-558(1998).
2www.nature.com/nature
doi: 10.1038/nature05457 SUPPLEMENTARY INFORMATION
10 Yin, C., Bengston, S., and Yue, Z., Silicified and phosphatized Tianzhushania,spheroidal microfossils of possible animal origin from the Neoproterozoic ofSouth China. Acta Palaeo. Polo. 49, 1-12 (2004).
11 Yin, L. and Li, Z., Precambrian microfloras of southwest China with reference totheir stratigraphic significance. Mem. Nanjing Inst. Geol. Palaeo., Acad. Sinica.10, 41-108 (1978).
12 Zhang, Y., Yin, L., Xiao, S., and Knoll, A. H., Permineralized fossils from theTerminal Proterozoic Doushantuo Formation, South China. Paleo. Soc. Mem. 50,52 p. (1998).
13 Xue, Y. S., Tang, T. F., and Yu, C. L., "Animal embryos," a misinterpretation ofNeoproterozoic microfossils. Acta Micropalaeo. Sinica 16, 1-4 (1999).
14 Xiao, S. and Knoll, A. H., Embryos or algae? A reply. Acta Micropalaeo. Sinica16, 313-323 (1999).
15 Xiao, S. and Knoll, A. H., Fossil preservation in the Neoproterozoic Doushantuophosphorite Lagerstätte, South China. Lethaia 32, 219-240 (1999).
16 Dornbos, S.Q. et al., Environmental controls on the taphonomy of phosphatizedanimal and animal embryos from the Neoproterozoic Doushantuo Formation,Southwest China. Palaios 21, 3-14 (2006).
2www.nature.com/nature
doi: 10.1038/nature05457 SUPPLEMENTARY INFORMATION