- The Paleontological Society
Multicellular filaments from the ca. 1200-Ma Hunting Formation (Somerset Island, arctic Canada) are identified as bangiacean red algae on the basis of diagnostic cell-division patterns. As the oldest taxonomically resolved eukaryote on record Bangiomorpha pubescens n. gen. n. sp. provides a key datum point for constraining protistan phylogeny. Combined with an increasingly resolved record of other Proterozoic eukaryotes, these fossils mark the onset of a major protistan radiation near the Mesoproterozoic/Neoproterozoic boundary.
Differential spore/gamete formation shows Bangiomorpha pubescens to have been sexually reproducing, the oldest reported occurrence in the fossil record. Sex was critical for the subsequent success of eukaryotes, not so much for the advantages of genetic recombination, but because it allowed for complex multicellularity. The selective advantages of complex multicellularity are considered sufficient for it to have arisen immediately following the appearance of sexual reproduction. As such, the most reliable proxy for the first appearance of sex will be the first stratigraphic occurrence of complex multicellularity.
Bangiomorpha pubescens is the first occurrence of complex multicellularity in the fossil record. A differentiated basal holdfast structure allowed for positive substrate attachment and thus the selective advantages of vertical orientation; i.e., an early example of ecological tiering. More generally, eukaryotic multicellularity is the innovation that established organismal morphology as a significant factor in the evolutionary process. As complex eukaryotes modified, and created entirely novel, environments, their inherent capacity for reciprocal morphological adaptation, gave rise to the “biological environment” of directional evolution and “progress.” The evolution of sex, as a proximal cause of complex multicellularity, may thus account for the Mesoproterozoic/Neoproterozoic radiation of eukaryotes.
The red algae are a study in extremes. Morphologically more diverse than any other group of algae, they range from single cells to large ornate multicellular plants (Woelkerling 1990). Uniquely among (nonfungal) eukaryotes they lack both flagella and centrioles, and any evidence of ever having had them (Pueschel 1990). They show greater molecular sequence divergence than the kingdom Fungi or the chlorophytes plus green plants (Ragan et al. 1994), and exhibit a remarkable, often bizarre range of reproductive strategies (Hawkes 1990; Hommersand and Fredericq 1990). Some species are capable of withstanding extreme environments, not least Bangia atropurpurea, which ranges from lacustrine to fully marine habitats (Geesink 1973; Sheath and Cole 1980). Their chloroplasts appear to be extremely primitive, having the same pigment complexes and unstacked thylakoids as cyanobacteria (Pueschel 1990); the chloroplast of Porphya purpureum retains more genes (and is therefore presumably that much more primitive) than that of any other known eukaryote (Reith 1995). Molecular phylogenetic analyses further suggest the red algae occupy a relatively basal position among extant eukaryotes (Stiller and Hall 1997). In addition to all this, the oldest taxonomically resolved fossil eukaryote is a red alga, specifically a 1.2-billion-year-old filamentous bangiophyte from arctic Canada. In all but detail, this fossil is indistinguishable from modern Bangia (Butterfield et al. 1990) and thus stands as a key datum point for considering early eukaryotic evolution. Here I formally describe this fossil, Bangiomorpha pubescens n. gen., n. sp., and consider its implications for understanding the evolution of sex, multicellularity, and the Mesoproterozoic/Neoproterozoic radiation of eukaryotes.
The fossils occur in the Hunting Formation, the upper unit of a small Proterozoic inlier that crops out on northwestern Somerset Island, arctic Canada (Stewart 1987) (Fig. 1). These little-altered shallow-water carbonates were part of a regional sedimentary basin that includes late Mesoproterozoic units on nearby Baffin Island and Greenland. The correlation is recognized on the basis of detailed stable isotope chemostratigraphy (Kah et al. 1999), as well as closely comparable litho- and biostratigraphy (see Hofmann and Jackson 1991). Pb-Pb dating of carbonates from the Hunting-correlative Society Cliffs Formation on Baffin Island has recently yielded a well constrained age of 1198 ± 24 Ma (L. Kah personal communication 1999), a major refinement on the previous radiometric determinations, which bracketed the Hunting Formation to between 1267 ± 2 Ma and 723 ± 3 Ma.
The 1000+ meters of the Hunting Formation are characterized by locally abundant microdigitate stromatolites, decimeter-scale fibro-radiate tufa(?), mud-cracks, tepee structures, and often pervasive diagenetic chert—all indications of a shallow-water to emergent intertidal/supratidal environment. The fossils reported here are from a ca. two-meter-thick section, low in the formation (estimated 100 m above the base), at the head of a secondary tributary to the Hunting River (Fig. 1). The exposure is a gray, laminated to thin-bedded dolostone with relatively continuous layers of dark-gray to black chert up to several centimeters thick. In thin-section, the dark chert layers are seen to include two sediment types: (1) faintly laminated fine sediment (originally micrite?) containing streaks of compacted organic material but few fossils, and (2) discrete but irregular laminations ranging from a few tens of micrometers to several millimeters thickness and often abundantly fossiliferous. The tendency of this latter lithology to form large (cm-sized) tabular intraclasts shows them to have been early stabilized/mineralized crusts; the common clustering of Bangiomorpha colonies on these surfaces (Fig. 2) points to an ecological preference for firm substrates.
Petrographic thin-sections of these silicified carbonates reveal an abundance of exceptionally well-preserved fossils. Of the six locally derived samples containing fossils, five were dominated by associations of Bangiomorpha n. gen. and the stalk-forming cyanobacterium Polybessurus (Green et al. 1987). Subordinate forms include both simple (Myxococcoides) and sheathed spheroids (Gloeodiniopsis spp.,? Pterospermopsimorpha), large bilayered filaments (Rugosoopsis), and two previously undocumented forms that may represent additional multicellular eukaryotes (Butterfield in press). Siphonophycus, the (almost) ubiquitous filamentous microfossil of Proterozoic microbial mats, is conspicuously absent from all of the Bangiomorpha-bearing samples.
All type and illustrated fossil specimens are housed in the Paleobotanical Collections of Harvard University (HUPC). Individual specimens, all in petrographic thin-section, are located using England Finder coordinates.
Domain Eucarya Woese, Kandler, and Wheelis, 1990
Division Rhodophyta Wettstein, 1924
Class Bangiophyceae Melchior, 1954
Order Bangiales Schmitz, 1892
Family Bangiaceae Nägeli, 1847
Genus Bangiomorpha n. gen
With reference to the marked similarity with modern Bangia.
Bangiomorpha pubescens n. sp.
Unbranched, vertically oriented filamentous bangiaceans with a basal multicellular holdfast structure. Individual cells defined by thin, relatively dark walls; whole plant enclosed in a thick translucent layer. Filaments uniseriate, multiseriate, or a combination of the two; positioning of cells in multiseriate portions reflects their derivation from uniserial precursors. Uniseriate filaments constructed of stacked disk-shaped cells, paired hierarchically into groups of two, four, and occasionally eight cells. Multiseriate portions of filaments generally constructed of four to eight radially oriented wedge-shaped cells; alternatively, of relatively few isolated spheroidal cells, or many close-packed spheroidal cells.
Bangiomorpha n. gen. is in all substantial respects indistinguishable from the modern bangiophyte red alga Bangia (Table 1). The key diagnostic character (synapomorphy) establishing its bangiacean affinity is the fourfold radially symmetrical arrangement of wedge-shaped cells that constitute most multiseriate filaments; this habit records the unique pattern of longitudinal intercalary cell division that is otherwise known only in modern Bangia. Likewise, the hierarchical pairing of cells in uniseriate filaments documents the bangiacean habit of diffuse growth whereby all vegetative cells contribute to initial filament elongation through transverse intercalary cell division (vs. the apical growth of most other algae and filamentous cyanobacteria).
Despite these marked similarities, the first issue must be taxonomic. Can Bangiomorpha be assigned unambiguously to the bangiacean red algae? Among cyanobacteria, the default assignment for Proterozoic microfossils, the only reasonable comparison is with the filamentous Stigonematales. Most Stigonematales differentiate a comparably thick external sheath, and many are multiseriate. They are also, however, characterized by true branching, apical growth, and differentiated heterocysts and akinetes (Martin and Wyatt 1974; Anagnostidis and Komárek 1990), none of which is found in Bangiomorpha. Furthermore, no stigonematalean, or indeed any known cyanobacterium, exhibits the regular radial intercalary cell division seen in Bangia/Bangiomorpha (Table 1). This latter feature likewise rules out comparison with various eukaryotic filaments, e.g., species of the green algal family Schizomeridaceae (filamentous Ulotrichales) (Table 1) or more primitive filamentous bangiophyceans belonging to the Porphyridiales (e.g., Goniotrichum) or Erythropeltidales (Erythrotrichia). Intercalary radial division has been observed in the modern prasiolalean chlorophyte Rosenvingiella; however, it does not show the regular fourfold symmetry of filamentous bangiaceans (Table 1) (Scagel 1966: Plate 13D). More importantly, the radial cells of Rosenvingiella are decidedly transient as they proceed to divide in three planes producing the solid parenchymatous spheres characteristic of the genus.
The fossils differ from modern Bangia in a number of ways, the most obvious being the possession of a multicellular holdfast structure; by contrast, modern Bangia typically has elongated, nonseptate rhizoids descending from a number of basal vegetative cells (Sommerfeld and Nichols 1970: Fig. 3). Geesink (1973), however, has reported holdfast structures in Bangia that are limited to the basal cell, and other filamentous bangiophytes, though not Bangiales, do have multicellular holdfasts closely comparable to those of Bangiomorpha (e.g., Erythrotrichia [Garbary et al. 1980: Fig. 4n]). Insofar as holdfasts are related critically to the local turbulence and substrate conditions, their particular form might be considered relatively plastic (though this is not always the case [C. Pueschel personal communication 1999]). Accessory, but less compelling, reasons for separating Bangiomorpha from Bangia are the 1.2-billion-year difference in their ages and the relative timing of holdfast initiation (see below).
Enzien (1990) has described a short uniseriate filament fragment with evidence of transverse intercalary cell division (paired disk-shaped cells within an outer sheath) from the Narssârssuk Formation, northwestern Greenland, and Kah and Knoll (1996) have mentioned similar material from the Society Cliffs Formation on adjacent Baffin Island. Both of these units are at least broadly correlative with the Hunting Formation (Kah et al. 1999) and represent similar peritidal carbonate environments. The possibility that the Narssârssuk and Society Cliffs fossils are also Bangiomorpha is intriguing, but in the absence of the diagnostic multiseriate filaments or a holdfast structure they cannot be distinguished unambiguously from a Johannesbaptistia-like cyanobacterium. Sheathed multiseriate filaments comparable in size to Bangiomorpha are characteristic of Palaeomicrocystis schopfii Maithy, 1975 from the Neoproterozoic Bushimay System of Zaire/Congo; however, the flattened nature of these compression fossils prevents analysis of the cell shape and division patterns, and therefore their higher-order taxonomy.
Bangiomorpha pubescens n. sp. (Figs. 2–6)
With reference to its pubescent or hairlike form, as well as the connotations of having achieved sexual maturity.
HUPC 62912, Figure 5E, Slide HUST-1A, England Finder coordinates: O-35.
Lower Hunting Formation, Somerset Island, arctic Canada. Field locality 87-HUST; 73°35.5‘N,94°46‘W.
A species of Bangiomorpha with uniseriate portions of filaments less than 50 μm wide.
Vertically oriented uniseriate and multiseriate unbranched filaments up to 2 mm long; isolated or gregarious in clusters of up to 15 individuals; commonly colonizing localized firm substrates. Uniseriate forms 15–45 μm wide (x̄ = 25.4 ± 6.0 μm; n = 500) with disk-shaped cells hierarchically paired and/or with a circumferential furrow recording incipient centripetal division. Multiseriate forms 30–67 μm wide (x̄ = 45.7 ± 8.6 μm; n = 27), usually with four to eight radially arranged wedge-shaped cells present in transverse cross-section. Multiseriate filaments occasionally constructed of four spheroidal cells isolated by translucent outer wall material or, rarely, of many close-packed spheroidal cells with no intervening matrix. Some filaments with intervals of both the uniseriate and multiseriate habits, the ratio of multiseriate to uniseriate filament diameter ranging from 1.0 to 2.1 (x̄ = 1.5 ± 0.3; n = 13). All filaments surrounded by a prominent but relatively translucent outer wall. Lobate multicellular holdfast structure connected to remainder of filament via a single cell.
Chert/carbonate sample 87-HUST-1 (20 thin-sections, 1000+ specimens); 87-HUST-2 (2 thin-sections); 87-HUST-5 (5 thin-sections); 87-HUST-7 (3 thin-sections); 87-HU-ST-8 (2 thin-sections).
Thin-section number and England Finder coordinates in parentheses. HUPC 62995, Figure 3B (HUST-1Q, O-45); HUPC 62996, Figure 4E (HUST-1E, J-25); HUPC 62997, Figure 4F (HUST-1A, N-40); HUPC 62998, Figure 5C (HUST-1R, M-17); HUPC 62999, Figure 5E (HUST-1A, O-34); HUPC 62919, Figure 5H (HUST-1B, M-42); HUPC 62910, Figure 5I (HUST-1A, O-37); HUPC 63011, Figure 6A (HUST-1C, N-38).
Most multicellular organisms undergo an ontogenetic development originating from a single cell. This certainly was the case with Bangiomorpha pubescens n. gen., n. sp., and the large populations in the Hunting Formation allow a nearly complete reconstruction of its ontogeny (Figs. 3–5). Although diagnosed on its multicellular habit, the initial single-celled (Fig. 4A) and double-celled (Fig. 4B) stages of Bangiomorpha can be identified by the specific character of their cell walls, in particular the relatively dark, pointillistically textured inner cell wall surrounded by a relatively translucent outer wall. Filament growth was initiated by the first cell division, oriented parallel to the substrate. By the four-celled stage (Fig. 4C), the characteristic pairing of cells reveals the transverse intercalary nature of cell division in uniseriate filaments; centripetal cytokinesis is documented by the common occurrence of prominent circumferential furrows (e.g., Fig. 3B). The basal holdfast is first seen to differentiate at the ca. 12–16 cell stage (Fig. 4F,G) and typically develops as a multilobed (usually two, but sometimes four or more) multicellular structure connected to the rest of the filament via a single cell (Fig. 6).
At some, presumably relatively mature, stage the cells of some Bangiomorpha filaments underwent longitudinal (with respect to the filament) intercalary division giving rise to multiseriate filaments. There are, however, at least three variations to the general pattern: Type 1: In most instances the intercalary division was oriented radially resulting in four or eight wedge-shaped cells arranged around a central space (Fig. 5E). The mean diameter of all such filaments is 46.2 ± 7.4 μm (n = 23). In specimens with both uniseriate and multiseriate portions (n = 9), the mean multiseriate diameter is 42.0 ± 5.5 μm and the adjacent uniseriate diameter 30.6 ± 6.4 μm; the ratio of uniseriate to multiseriate diameter ranges from unity to 1.8 (x̄ = 1.4 ± 0.3). Type 2: In a few instances, longitudinal intercalary division gave rise to relatively few spheroidal cells separated from one another by translucent outer wall material (Fig. 5A,D). Mean filament diameter is 40.0 ± 9.1 μm (n = 4). In specimens with both uniseriate and multiseriate portions (n = 3), the mean multiseriate diameter is 36.7 ± 8.0 μm and the adjacent uniseriate diameter 24.0 ± 7.0 μm; the ratio of uniseriate to multiseriate diameter ranges from 1.2 to 2.1 (x̄ = 1.6 ± 0.4). Type 3: In a single specimen, the multiseriate portion of a filament is composed of many close-packed spheroids with translucent outer-wall material limited to the margins of the filament (Fig. 5H); it is 63 μm in diameter and the adjacent uniseriate portion 30 μm, for a ratio of 2.1. This specimen also includes a basal holdfast, thus revealing the basipetal maturation of Bangiomorpha.
Clearly there is a wide range of variation in the multicellular habit of Bangiomorpha. In some instances, for example, the shift to the multiseriate condition was accompanied by considerable filament expansion (Fig. 5A–C,H), and in others none (Fig. 3C). Although most multiseriate filaments are associated with relatively large-diameter uniseriate filaments, this is not always the case (e.g., Fig. 5A). There is also an indication (not statistically significant) that Type 2 multiseriate filaments were narrower and derived from somewhat smaller-diameter uniseriate filaments than the other two types. In any event, with uniseriate filaments up to 45 μm diameter and multiseriate filaments as narrow as 30 μm, it is clear that maturation in Bangiomorpha was not simply a matter of size. With ecophenotypic effects limited by the close proximity of all these filaments (all types occur in a single hand sample), it is clear that Bangiomorpha had at least two, possibly more, distinct ontogenetic fates.
Comparison with Modern Bangia
In most details of its morphology and ontogeny, Bangiomorpha n. gen. compares closely with the haploid phase of the modern red alga Bangia, a Recent cosmopolitan seaweed that colonizes shallow-water, often emergent, hard substrates in both marine and freshwater settings (Geesink 1973; Sheath and Cole 1980, 1984; Sheath et al. 1985). Modern Bangia has a biphasic life cycle with a macroscopic, haploid, gametophytic generation—the “bangia” phase—alternating with a microscopic, diploid, sporophytic “conchocelis” phase. Bangia-phase filaments are unbranched and, when immature, composed of a single (uniseriate) row of stacked, disk-shaped cells. The polysaccharide cell walls are distinctly biphasic, with a conspicuous inner wall that defines individual cells and an outer wall enveloping the whole organism (Cole et al. 1985); the translucent material that constitutes this outer wall may also occupy the central part of more mature multiseriate filaments (Fig. 7). A submicron-thick “cuticle” gives the filament a sharply delineated outer wall.
Gametophytic Bangia is notable in having three reproductive types—asexual, male, and female—which can be distinguished on the basis of cellular morphology. Female Bangia filaments tend to be the largest diameter, males intermediate and asexual filaments the narrowest, although absolute dimensions vary between habitats and regions. For example, populations of asexual filaments have a mean maximum diameter of 75.3 ± 0.9 μm in the Great Lakes, but 157 ± 3.2 μm in the Pacific Ocean; males range from 86.2 ± 4.1 μm in the Atlantic to 124 ± 7.6 μm in the Pacific; and females from 127 ± 4.0 μm in the Atlantic to 171 ± 8.1 μm in the Pacific (Sheath and Cole 1984). Germlings, however, are considerably narrower, 11.1–20.8 μm diameter (Sheath and Cole 1984), and Geesink (1973) has reported filaments from the Netherlands up to 15 cm long that measure just ca. 20 μm diameter at the (uniseriate) base and 20–80 μm at the (multiseriate) apex.
Growth of a Bangia filament from a germinating asexual monospore (or a conchocelis-derived conchospore) begins with the differentiation of a basal attachment structure followed by cell division parallel to the substrate (Sommerfeld and Nichols 1970) (unlike in Bangiomorpha, where holdfast differentiation is initiated only at the 12–16 cell stage). Except for this basal cell, all cells then divide in the same plane (diffuse intercalary cell division) to produce an unbranched uniseriate filament.
Uniseriate Bangia filaments are induced to become multiseriate by specific light and temperature conditions (Sommerfeld and Nichols 1973), a process that begins at the apex and proceeds basipetally. The first stage involves a longitudinal intercalary cell division of a distal uniserial cell, resulting in 4, 8, or 16 wedge-shaped cells arranged radially around a central space (Fig. 7A–C). Depending on the filament type, the diameter of multiserial portions of filaments may be either several times that of uniserial portions or much the same (Sommerfeld and Nichols 1970). In either case, the resultant daughter cells typically retain their positions and thereby reflect the original uniserial arrangement of the mother cells.
Subsequent maturation involves the production of reproductive spores, which can be recognized by (1) further intercalary division, (2) a spatial separation of the constituent cells, and/or (3) the acquisition of spheroidal cell shape. It is at this stage that it is possible to distinguish, on the basis of morphology, asexual, male, and female filaments. In asexual Bangia, individual vegetative cells are transformed directly into asexual reproductive spores without intervening division (Garbary et al. 1980; but see Cole et al. 1985: Fig. 11 for a possible exception); thus a single uniserial cell, having divided radially to produce 4–16 vegetative wedge-shaped cells, will produce 4–16 monospores. In male Bangia, each of the wedge-shaped cells undergoes repeated divisions to produce up to 128 small, colorless spermatia; thus an original uniserial cell might yield up to 2048 spermatia (Fig. 7E). In female Bangia, the wedge-shaped cells transform directly into carpogonia, which, following fertilization, divide to form 8–16 carpospores (Fig. 7D); thus an original uniserial cell typically yields 4–16 carpogonia, but 32–256 carpospores.
Released carpospores germinate to produce the diploid conchocelis phase of the Bangia life cycle. The conchocelis differs conspicuously from the bangia phase, having a microscopic, branching, uniseriate thallus that grows endolithically within carbonate substrates (Cole and Conway 1980). Following meiotic reduction, the conchocelis produces haploid conchospores, which complete the life cycle by germinating bangia-phase filaments. The bangiophyte conchocelis is thought to be of particular evolutionary importance in that it expresses a variety of characteristics typical of the florideophyte red algae, e.g., apical growth, plugged pit connections, multiple ribbon-shaped chloroplasts, peripherally positioned chloroplasts, and a central vacuole (Cole and Conway 1975). The implication is that the florideophytes may have arisen as a consequence of sexual reproduction in bangiophytes.
Given the compelling case for identifying Bangiomorpha as a bangiacean red alga, it is worth considering the marked variation in (multiseriate) form as reflecting different reproductive types, as it does in modern Bangia. It is the mature, spore-bearing filaments (i.e., those bearing differentiated spheroidal cells) that provide the most convincing evidence: whereas Type 2 filaments (Fig. 5A,D) appear to have derived just a single spore from each of four vegetative, wedge-shaped cells, Type 3 filaments (Fig. 5H) clearly underwent significant tertiary division such that each wedge-shaped cell produced many spores. By comparison with Bangia, the former would appear to represent asexual monospores or unfertilized carpogonia, and the latter fertilized carpospores or spermatia (compare with Fig. 7) (Garbary et al. 1980). By the same token, Type 1 filaments are identified as vegetative plants that have yet to differentiate identifiable spore or gamete types. Although imperfect preservation frustrates exact identification of particular reproductive types in Bangiomorpha, the marked differences in ontogenetic pattern make a convincing case for the presence of differing reproductive types and therefore sexual reproduction. Positive identification of sex in the Proterozoic is of course significant in that it has been regularly invoked (Schopf et al. 1973; Stanley 1975; Knoll 1992), but rarely documented (see Zhang and Yuan 1996), as a key innovation in early eukaryotic evolution. Bangiomorpha is the earliest direct evidence for such a habit.
There is no direct evidence for a conchocelis phase—the expected consequence of bangiophyte sexual reproduction—in the Hunting assemblages. This is not entirely unexpected given the distinct habits (and accompanying taphonomies) of the two phases. The oldest, indeed the only, reported fossil conchocelis occurs in Late Silurian rocks from Poland (Campbell 1980) and is not associated with a haploid phase. It may be, of course, that a conchocelis does occur in the Hunting assemblages but has simply not being recognized as such; certainly there is no necessity that the entire life cycle of Bangiomorpha mirror that of modern Bangia. In any event, absence of a recognized conchocelis in no way implies that Bangiomorpha was asexual (see below).
THE EARLY FOSSIL RECORD OF EUKARYOTES
With a well-constrained age of ca. 1200 million years, Bangiomorpha is the oldest taxonomically resolved eukaryote yet reported and thus represents a key (minimum) datum point for constraining eukaryote evolution. Certainly there are older eukaryotic fossils, but a dearth of preserved characters precludes any satisfactory taxonomic placement. Macroscopic coiled filaments from the Paleoproterozoic Negaunee Iron Formation, Michigan, for example, have been compared with large coenocytic chlorophytes (Han and Runnegar 1992), but realistically cannot be distinguished from numerous other groups, including colonial prokaryotes. The same is true for the diverse macrofossils reported from the ca. 1700-Ma Chuanlinggou and Tuanshanzi Formation s (Changcheng System), Jixian, China (Zhu and Chen 1995; Yan 1995; Yan and Liu 1997); an exception here is Qingshania (Chuanlinggou Fm.), a relatively large-diameter filament which is at least convincingly eukaryotic (Yan 1989). Early eukaryotes are also recognized among the acritarchs beginning around 1850 Ma (Zhang 1997). Although of unknown taxonomic affinity, relatively large acritarchs (>50 μm diameter) are generally accepted as representing a eukaryotic grade of organization. Eukaryotic biomarker molecules are now recognized as far back as the late Archean (2.7 Ga [Brocks et al. 1999]).
Eukaryote diversity remains conspicuously low through most of the Mesoproterozoic (1600–1000 Ma). Larger acritarchs, although widespread, are represented by simple sphaeromorphs (Knoll 1994). The coiled macrofossil Grypania, reasonably interpreted as an individual eukaryote and broadly distributed by ca. 1400 Ma (Walter et al. 1990; Kumar 1995), appears to be unaccompanied by other multicellular/macroscopic fossils (with the possible exception of problematic “string of beads” imprints from Montana and Australia [Horodyski 1993]). It is only toward the end of the Mesoproterozoic that diversity is seen to increase, most notably perhaps with the appearance of Bangiomorpha (Rhodophyta) and other possible eukaryotes in the Hunting Formation (Butterfield in press). In the terminal Mesoproterozoic and/or earliest Neoproterozoic, diversity rises further with the introduction of increasingly diverse ornamented acritarchs (Knoll 1992, 1994; Xiao et al. 1997) and, in the ca. 1000-Ma Lakhanda Formation of Siberia, the first appearance of a fossil stramenopile, Palaeovaucheria (Hermann 1981; Woods et al. 1998). Diversity and taxonomic resolution continue to rise through the early–middle Neoproterozoic with the identification of possible dinoflagellates (alveolates) in the ca. 850-Ma Wynniatt Formation, arctic Canada (Butterfield and Rainbird 1998) and of several taxa of green algae (chlorophytes) in the ca. 750-Ma Svanbergfjellet Formation, Spitsbergen (Butterfield et al. 1994). Possible “phylloid” algae are also reported from the early Neoproterozoic Pahrump Group of California (Horodyski and Mankiewicz 1990); “scale microfossils” from the middle Neoproterozoic Tindir Group, Alaska, are reminiscent of certain chrysophyte or diatom elements (Allison and Hilgert 1986; Kaufman et al. 1992); and vase-shaped microfossils found through much of the Neoproterozoic are interpreted as testate amoebae (Porter and Knoll, this issue). Thus, by the middle of the Neoproterozoic there is good fossil evidence for the divergence of most of the major algal/protistan clades.
Despite the early appearance of these sundry groups, it is important to recognize that the data are exceedingly sparse, typically single occurrences that precede the next appearance in the fossil record by hundreds of millions of years. Taken individually they are unlikely to offer even a crude approximation of actual first appearances. Collectively, however, there does appear to be a genuine signal, with the interval between 1200 and 700 Ma conspicuously more diverse than the preceding 500 million years. The reality of this pattern—to the extent of documenting a major Meso/Neoproterozoic radiation of eukaryotes (Knoll 1992)—is bolstered by the accompanying diversity increase among the much better sampled acritarchs (Knoll 1994), as well as molecular analyses suggesting a “big bang” origination of the principal eukaryotic lineages (Knoll 1992; Philippe and Adoutte 1998; Philippe et al. in press). Such an interpretation accords with the pulsed pattern of diversification expected during adaptive radiation, though there is no evidence that this proceeded at a rate comparable to, say, the “Cambrian explosion”; indeed, evidence from the acritarch record points to a fundamentally slower evolutionary turnover among Proterozoic eukaryotes compared to their Phanerozoic counterparts (Knoll 1994).
THE RED ALGAE IN EARLY EUKARYOTIC EVOLUTION
The red algae have been shunted widely about in eukaryotic phylogenies. Beginning in the mid–nineteenth century (see Ragan and Gutell 1995), they were considered among the most ancient eukaryotes because of their universal lack of flagella, basal bodies and centrioles and their conspicuously cyanobacteria-like chloroplasts (e.g., phycobilin pigments and unstacked thylakoids). This basal positioning acquired some support from early molecular phylogenetic analyses based on 5S rRNA (Hori and Osawa 1987). Subsequent work, however, using the larger and presumably more reliable SSU rRNA (Bhattacharya et al. 1990; Hendriks et al. 1991; Kumar and Rzhetsky 1996) and LSU rRNA (Perasso et al. 1989), positioned them relatively late in eukaryotic evolution, arising more or less coincidentally with other major eukaryotic clades; e.g., the alveolates, stramenopiles, chlorophytes, fungi, and metazoans. Within these “crown eukaryotes” the red algae have been considered a sister group of the chlorophytes (Ragan and Gutell 1995) or of a more basal plant/animal/fungi clade (Kumar and Rzhetsky (1996), despite the widespread recognition that rRNA offers limited phylogenetic resolution at this level (Ragan and Gutell 1995; Delwiche and Palmer 1997; Hirt et al. 1999; Phillippe et al. in press).
Most other molecules that have been applied to eukaryotic phylogeny (e.g., elongation factors, tubulins, hsp70, GAPDH) yield similarly tenuous results (Delwiche and Palmer 1997). By contrast, RPB1, the gene encoding the largest subunit of RNA polymerase II, provides strong statistical support for the divergence of the Rhodophyta diverging before the last common ancestor of an unresolved “crown” (Stiller and Hall 1997), indeed one of the few instances of statistical support for any relationship at this taxonomic level (Delwiche and Palmer 1997; but see Moreira et al. 2000). Insofar as many, perhaps all (Philippe et al. in press) of the purportedly ancient eukaryotic groups that occupy the “stem” of SSU rRNA–based trees are now recognized as derived constituents of the “crown” (Hirt et al. 1999; Stiller and Hall 1999; Philippe et al. in press), the red algae might once again be considered an early-diverging lineage. By extension, their lack of flagella and related structures might thus reflect the ancestral eukaryotic condition.
Within the red algae, molecular analyses have contributed more clearly to phylogenetic resolution. The two rhodophyte classes (Bangiophyceae and Florideophyceae) are separated on the basis of both SSU rRNA (Ragan et al. 1994) and plastid rbcL (Freshwater et al. 1994), with the bangiophytes consistently occupying the most basal branches. Unlike the clearly monophyletic florideophytes, the bangiophytes are highly divergent, to the extent that they are likely to be “polyphyletic” (= paraphyletic [Ragan et al. 1994]). Within the Bangiophyceae, the order Bangiales appears to be relatively derived and monophyletic, though notably divergent, as with the red algae as a whole (Ragan et al. 1994).
By introducing large size, complex morphology, and thereby increasingly complex ecology, the introduction of eukaryotic multicellularity represents a critical threshold in the history of life. Such organization is first documented convincingly by the simple cellular filaments of ca. 1700-Ma Qingshania, followed by the macroscopic, possibly coenocytic Grypania at ca. 1400 Ma. Although not the oldest multicellular eukaryote, Bangiomorpha nevertheless holds special status as the first on record to exhibit true cellular differentiation and specialization. Unlike any of its predecessors, Bangiomorpha had, for example, a differentiated holdfast, multiple cycles of cell division, differentiated spores, and sexually differentiated whole plants.
In one sense, multicellularity is nothing more than the substitution of somatic for reproductive mitosis (Kondrashov 1997). In the simplest case this will result in a multicellular mass of identical cells. Differing local environments might then lead to differing fates for particular cell lineages and so to the emergence of (multicellular) organism-level characteristics. In the first instance this likely involved little more than intercellular attachment or arrangement, e.g., acquisition of a filamentous habit and/or vertical orientation. More profound organism-level characteristics would have arisen with the introduction of asymmetric cell division, where the differing fates of sister cells give rise to specialization and an intraorganismal division of labor (see Horvitz and Herskowitz 1992; Maynard Smith and Szathmáry 1995; Kirk 1998). Combined with the eukaryotic potential for large size and sophisticated development, such novelty would have opened up fundamental new areas of morphospace.
As with any evolutionary innovation, ecology assuredly played a critical role in the origin of eukaryotic multicellularity. Large size and complexity present a variety of otherwise unavailable habits conferring selective advantage, including predator evasion and an enhanced uptake and storage of essential nutrients (Kirk 1998). In the case of Bangiomorpha, and perhaps other early metaphytes, multicellularity also introduced a novel approach to benthic photosynthesis: with a differentiated basal attachment structure imparting current and wave resistance, multicellular filaments could now compete effectively for light and nutrients by growing vertically (Fig. 2). Vertical orientation in turn introduces a substantial new aspect to shallow-water environments, affecting bottom-water currents, sediment transport and trapping characteristics, chemical exchange, and a host of biological interactions (Carpenter and Williams 1993); indeed, vertically oriented turf-forming algae are recognized today as the basis of a distinct benthic community (Hay 1981). Insofar as modern algal turfs also contribute importantly to environmental patchiness and overall benthic heterogeneity (Airoldi and Virgilio 1998), it would also appear that the early evolution of vertical orientation—tiering—would have contributed importantly to the diversification of Proterozoic environments, just as it has in the Phanerozoic (e.g., Bottjer and Ausich 1986). Differential effects on sedimentation and sedimentary fabric no doubt played a further role in the mid-Mesoproterozoic diversification and early Neoproterozoic decline of stromatolites (see Grotzinger and Knoll 1999). More generally, by breaking the hegemony of horizontally oriented microbial mats, vertically oriented algal turfs might well have triggered a mutual feedback system of diversification (see Stanley 1973), broadly expressed as the Mesoproterozoic/Neoproterozoic radiation.
Other Mesoproterozoic organisms may also have contributed to this effect. Various bacteria and cyanobacteria are known to grow vertically, and simple vertically oriented filaments have previously been reported from late Mesoproterozoic microbial mat assemblages (Knoll and Sergeev 1995: Fig. 3). These, however, are on a much smaller scale than that seen with Bangiomorpha, and are limited fundamentally by their prokaryotic organization (see below). Mesoproterozoic Grypania has also been interpreted as a benthic macrophyte, presumably with the axis of its coiled filament oriented vertically (Walter et al. 1990; Runnegar 1994; Kumar 1995). It does not, however, show any evidence of a terminal holdfast or indeed any cellular or morphological differentiation.
The appearance of sexual reproduction—i.e., syngamy, genetic recombination and meiosis—has long been considered a major evolutionary threshold, giving rise to a fundamental increase in variation (Schopf et al. 1973), a novel ability to remove deleterious mutations (Muller 1964), and indeed “true” species and speciation (Stanley 1975). The presence of at least two distinct spore-producing phases in Bangiomorpha, and their close comparison to sexual phases in modern Bangia, presents a convincing case for eukaryotic sex by at least ca. 1200 Ma.
Sex appears to be plesiomorphic, if not necessarily monophyletic (Ruvinsky 1997), for all eukaryotes with a differentiated multicellular grade of organization (Bell 1982; Buss 1987; Grosberg and Strathman 1998; Dacks and Roger 1999), and there is a compelling case for considering sex to be a prerequisite for eukaryotic multicellularity. At one level, it would appear that the intercellular recognition and coordination critical to sexual reproduction are very much the same kinds of processes involved in the development of a multicellular organism. Thus, appearance of sex, for whatever reason, might have introduced a collateral capacity for multicellularity.
At another level, sex—in particular, obligate sex (Dacks and Roger 1999)—would appear to be necessary for the evolution of multicellularity simply to override the inherent conflict between the individual and its constituent cells (Buss 1987; Michod 1997). Differentiated, functionally specialized somatic cells are of obvious selective advantage to the whole organism but, by diminishing their individual reproductive capacity, are at a severe competitive disadvantage next to any cell retaining, or capable of reverting to, its original totipotent condition. In asexual lineages there is no way of excising such “somatic cell parasites” (Muller's ratchet), and any potential multicellular organism will, so it is argued, eventually reduce to a mass of undifferentiated cells. Sexual reproduction, by contrast, allows for the periodic removal of these “parasites,” allowing the selective forces on the multicellular individual to take precedence over that of its cells (Buss 1987). It is a powerful argument (despite the existence of cellularly differentiated prokaryotes, e.g., heterocystous cyanobacteria) and, insofar as Bangiomorpha features differentiated vegetative cells (e.g., the basal holdfast), corroborates its status as an early sexually reproducing organism. By contrast, neither Qingshania nor Grypania shows any evidence of the cellular differentiation that would suggest sexuality.
True somatic differentiation in Bangiomorpha was limited to its basal holdfast structure; apart from the differentiated spores/gametes, all other cells were effectively identical, underwent regular mitosis and contributed equally to plant elongation (i.e., diffuse intercalary division). This pattern is entirely distinct from the apical meristem–type growth of most (but not all [Table 1]) other metaphytes and would appear to represent a particularly primitive grade of multicellularity (progressive loss of cellular totipotency being the logical consequence of evolving multicellular lineages [see Buss 1987]). Indeed, the present fossil material might be interpreted as evidence for such a gradient even within the Bangiaceae: whereas holdfast differentiation in Bangiomorpha was not initiated, i.e., totipotency not abandoned, until the 12–16 cell stage (Fig. 4F,G), modern Bangia differentiates its basal cell following the first cell division.
Sexual reproduction results in a haploid-diploid alternation of generations, which itself offers a rich source of evolutionary novelty (Mable and Otto 1998). In the case of the red algae, the sexually derived diploid conchocelis phase of the Bangiales appears to be the link between the basal Bangiophyceae and the derived Florideophyceae (Cole and Conway 1975). The presence of sexually reproducing bangiophytes at ca. 1200 Ma thus suggests an early initiation of the florideophyte grade of organization, if not necessarily its cladogenetic independence. The first fossils reasonably interpreted as florideophytes follow the Hunting bangiophytes by some 600 million years; i.e., phosphatized cellular thalli of Paramecia and Thallophyca in the terminal Neoproterozoic Doushantuo Formation of South China (Zhang et al. 1998).
Bangiomorpha is the oldest taxonomically resolved eukaryote on record. As such it provides a key datum point for resolving/constraining protistan phylogeny, particularly molecular clock hypotheses (e.g., Kumar and Rzhetsky 1996). The precise taxonomic assignment of this fossil—to the family level—is based on detailed comparison with the living rhodophyte Bangia. Neither the morphology nor ontogeny of this lineage has changed appreciably for some 1200 million years. Such “living fossil”-type longevity might be ascribed to a variety of factors but is most likely related to the unusually harsh conditions of the upper intertidal habitat; like Bangia, Bangiomorpha would have been regularly exposed to highly variable salinities and intervals of complete desiccation. It is interesting to note here that other early multicellular fossils also exhibit pronounced morphological stasis and likewise have modern counterparts with broad salinity tolerances; e.g., both Vaucheria (cf. ca. 1000-Ma Palaeovaucheria Hermann 1981) and Cladophora (cf. ca. 750-Ma Proterocladus Butterfield 1994) range from marine to freshwater and even subaerial environments (Fritsch 1935).
With its clear differentiation of sexual morphs Bangiomorpha is the oldest confirmed example of eukaryotic sex. There is of course, a much more ancient record of eukaryotes (Brocks et al. 1999), as well as a good case for sex representing the plesiomorphic condition among exant eukaryotes (Dacks and Roger 1999). This is not to suggest, however, that sexual reproduction was coincident with the origin of the eukaryotic cell (though it might well be implicated in the much later radiation of crown-group eukaryotes [i.e., the last common ancestor of extant groups plus all its descendants]). On the other hand, there is a strong case to be made for a near-synchronous appearance of sexual reproduction and complex multicellularity: in addition to sex being necessary for a differentiated multicellular grade of organization (through its ability to shed somatic cell parasites), the clear ecological advantages of eukaryotic multicellularity virtually assure its immediate succession. If this were the case, then the most reliable proxy for the appearance of sexual reproduction would be the first appearance of complex multicellularity. Bangiomorpha in the ca. 1200-Ma Hunting Formation is the oldest confirmed example of complex multicellularity.
The existence of numerous entirely unicellular eukaryotic clades suggests that complex multicellularity was not the primary impetus for the evolution of sex (see Bell 1982). Rather, sex introduced the opportunity of multicellularity, which was then exploited in various lineages and in various ways as selective advantages were discovered. Although limited initially, such opportunities would have expanded as accumulating morphology created novel environments and increasingly complex ecologies. Intriguingly, Stiller and Hall (1998) have presented molecular evidence for the particularly primitive nature of multicellularity in the red algae.
Both the fossil and molecular records point to an early but ecologically inconspicuous history of eukaryotic life, followed by a major radiation somewhere in the vicinity of the Mesoproterozoic/Neoproterozoic boundary (1 Ga). What was responsible for this “big bang of eukaryotic evolution”? Or, alternatively, how was it that eukaryotes appeared so early but took so long to diversify (or indeed to make a significant contribution to the biogeochemical record [Summons et al. 1999])? Brasier and Lindsay (1998) argue that much of this time was occupied simply assembling the various components of the modern eukaryotic cell. The mitochondrion, however, is probably more closely tied to the Archean origin of eukaryotes (Martin and Müller 1998), and the chloroplast to their Paleoproterozoic appearance in the body-fossil record (Knoll 1992). Alternatively, delayed evolution/acquisition of the eukaryotic flagellar apparatus might be invoked, particularly in light of arguments for the aflagellate red algae representing the ancestral eukaryotic condition (see above).
The more common explanation for the Mesoproterozoic/Neoproterozoic radiation is that it followed from the appearance of sexual reproduction (Schopf et al. 1973; Knoll 1992). This, I think, is surely correct, but not for the reasons usually proffered. True, sex would have greatly enhanced the genetic flexibility of eukaryotes, but this alone cannot explain their expansion into a biosphere already monopolized by highly adaptable prokaryotes. Without recourse to mass extinction, the displacement of such incumbents requires clear-cut ecological superiority. The one, perhaps only, field in which eukaryotes categorically outpace the Bacteria and Archaea is in their capacity for morphological differentiation, multicellularity, and large size. The reason sex was critical in eukaryotic evolution was that it introduced organismal morphology as a significant evolutionary factor, a new way to play the game that was simply beyond the means of the prokaryotic grade of organization.
At one level this alone is sufficient—after all, complex multicellularity constitutes much of the phenomenon we are seeking to explain. The more fundamental consequence of multicellularity, however, has to do with the revolutionary effects of morphology on Proterozoic ecology. Whereas the developmental/morphological simplicity of prokaryotes condemns them forever to operate in a simplistic physical environment, multicellular eukaryotes invented a biological environment, in particular, one with a virtually unlimited capacity for morphological complexity and size. With such conditions, of course, comes a host of complex ecological interactions, including powerful feedback effects on environment (ecological engineering, sensu Jones et al. 1997), organismal coevolution, adaptive reciprocity, and “progress” (indeed, these are an underlying assumption of logistic models of diversification for the Phanerozoic [e.g., Sepkoski 1978]). The new world order was a product of multicellular eukaryotes, which in turn was a consequence of sexual reproduction. The Mesoproterozoic/Neoproterozoic radiation of eukaryotes might have begun quite trivially—perhaps tiering on a scale somewhat larger than possible with prokaryotes or asexual eukaryotes—but the discovery of progressive morphological adaptation was to lead, perhaps inevitably, to the modern world of large, complex organisms.
I thank R. Buick, J. Stiller, H. Philippe, and S. Conway Morris for useful discussion, and A. Knoll and C. Pueschel for very constructive reviews. Cambridge Earth Science Contribution 5841.
- Accepted 7 February 2000.