A case for an active eukaryotic marine biosphere during the Proterozoic era
Contributed by Donald E. Canfield; received December 26, 2021; accepted August 25, 2022; reviewed by Shuhai Xiao and Emily Zakem
Significance
Eukaryotes evolved at least 1,700 million years ago (Ma), and they dominate marine ecosystems today, driving carbon and nutrient cycles. However, it is thought that prokaryotes dominated marine ecosystems until around 780 Ma, beginning the so-called “rise of algae.” Still, the ancient microfossil record shows a relatively diverse collection of large microfossils, including eukaryotes, from about 1,700 Ma onward. We use a size- and trait-based ecosystem model to show that the preserved microfossil record is most compatible with an active eukaryote ecosystem conducting osmotrophy, photosynthesis, and phagotrophy. Thus, marine ecosystems from 1,700 to 780 Ma were far more active and diverse than generally thought.
Abstract
The microfossil record demonstrates the presence of eukaryotic organisms in the marine ecosystem by about 1,700 million years ago (Ma). Despite this, steranes, a biomarker indicator of eukaryotic organisms, do not appear in the rock record until about 780 Ma in what is known as the “rise of algae.” Before this, it is argued that eukaryotes were minor ecosystem members, with prokaryotes dominating both primary production and ecosystem dynamics. In this view, the rise of algae was possibly sparked by increased nutrient availability supplying the higher nutrient requirements of eukaryotic algae. Here, we challenge this view. We use a size-based ecosystem model to show that the size distribution of preserved eukaryotic microfossils from 1,700 Ma and onward required an active eukaryote ecosystem complete with phototrophy, osmotrophy, phagotrophy, and mixotrophy. Model results suggest that eukaryotes accounted for one-half or more of the living biomass, with eukaryotic algae contributing to about one-half of total marine primary production. These ecosystems lived with deep-water phosphate levels of at least 10% of modern levels. The general lack of steranes in the pre–780-Ma rock record could be a result of poor preservation.
Sign up for PNAS alerts.
Get alerts for new articles, or get an alert when an article is cited.
Both the micro- and macrofossil records point to the presence of eukaryotic organisms in marine ecosystems as early as 1,700 to 1,600 Ma (e.g., refs. 1–4), with many fossils displaying a combination of features that strongly favor eukaryotic, rather than prokaryotic, interpretation. For example, the microfossil Shuiyousphaeridium macroreticulatum is large (50 to 300 μm) and spherical, with a reticulated surface hosting numerous cylindrical processes radiating outward (5, 6). The outer wall is multilayered (7), and overall, the combination features, especially the complexity of the outer wall and associated processes, strongly support a eukaryote interpretation (1, 3). The earliest representative of S. macroreticulatum is found in the Ruyang Group in northern China (5, 6), with a maximum deposition age of 1,744 Ma (8) and a likely age of about 1,600 Ma (9). The microfossil Tappania plana, of similar age to S. macroreticulatum, is also relatively large (20 to 160 μm) with an asymmetrical distribution of heteromorphic processes on the surface of a body that is often irregular in shape, indicating a vegetative cell (1, 3, 10). This microfossil is widely distributed in time and space and is generally accepted as a eukaryote (e.g., refs. 1–3, 10, and 11). Perhaps the oldest accepted eukaryote microfossil is Valeria lophostriata from the ca. (approximately) 1,700-Ma Chancheng Supergroup, northern China (7, 12, 13). This fossil, with vesicle diameters of ca. 60 to 250 μm (14), displays 1-μm-spaced concentric striations on the inner surface of the outer wall, strongly suggesting a eukaryote affinity (1–3, 11, 14). There are many other possible eukaryote microfossil candidates in the 1,700 to 1,400 Ma time range (e.g., refs. 1–3 and 15–17), suggesting that numerous eukaryotic clades were present in the Late Paleoproterozoic to Early Mesoproterozoic oceans.
From an energetic standpoint, eukaryotes can live by photosynthesis, phagotrophy (grazing), osmotrophy (taking up dissolved organic carbon [DOC] and nutrients), or some combination of these (e.g., ref. 18). The last common ancestor of crown-group eukaryotes (LECA) is generally viewed as a phagotroph, having fed on prokaryotes in the environment (18–21). If the early microfossils described above reflect crown-group eukaryotes originating after LECA, then grazing by phagocytosis was an ecosystem function as early as 1,700 Ma. If these fossils represent stem-group eukaryotes, then their relationship to phagotrophy depends on the relationship between phagotrophy and eukaryogenesis, a timing that is unclear (22, 23).
Indeed, some (e.g., refs. 24 and 25) have suggested that LECA emerged later in geologic time, as late as 1,200 to 1,000 Ma. This timing is consistent with the idea that stem-group eukaryotes may have inherited their mitochondria from processes other than phagocytosis (e.g., refs. 26 and 27) and that phagocytosis came later, having possibly evolved in multiple eukaryotic lineages after LECA (24). If this was the timing of events, then stem-group eukaryotes were osmotrophs, with phagotrophy and algal photosynthesis (with phagocytosis required to incorporate the original cyanobacterium that became the chloroplast) emerging after LECA. A prediction from this model is that eukaryote fossils before around 1,200 to 1,000 Ma were osmotrophic organisms.
The biomarker record of bitumen extracted from ancient sedimentary rocks might be consistent with a late emergence of LECA and of algal photosynthesis. Indeed, this record does not show the presence of eukaryotic algal steranes until about 780 Ma, beginning what has been called the “rise of algae” (28–30). In addition, the sterane to hopane ratio from extracted bitumens shows a large increase around 640 Ma (28). This ratio is taken to indicate a rise in the relative contribution of eukaryotes to prokaryotes in marine ecosystems. In addition, the biomarker 24-n-propylcholestane becomes apparent in the rock record concurrent with the rise in the sterane to hopane ratio (28). This biomarker is believed to be sourced from rhizarians, a predatory group of protists, including foraminifera and radiolarians, that actively capture their prey and ingest them through phagocytosis. Thus, the late emergence of the rhizaria might also support an overall late emergence of LECA.
There are other explanations, however, for the relatively late rise of algae and the apparent shift to a eukaryote-dominated ecosystem. These explanations include a Neoproterozoic-timed increase in atmospheric oxygen, releasing eukaryotes from inhibitory low oxygen concentrations (29, 31); the emergence of size-selective grazing of prokaryotes by protist predators (32, 33); and an increase in nutrient availability, with an emphasis on phosphorus (28, 34). In this idea, phosphate was actively scavenged by ferric iron formed at the chemocline of deep anoxic ferruginous waters during Mesoproterozoic times. In this way, the marine phosphorus inventory was reduced to concentrations much lower than today (34, 35). Due to their higher surface area to volume ratios, smaller cells are superior competitors for nutrients. Thus, reduced P concentrations would have benefited smaller classes of phytoplankton, such as cyanobacteria (36–38). Subsequently, massive Neoproterozoic era glaciations (“Snowball Earth” events) enhanced continental weathering, releasing large quantities of P into the oceans and providing a nutrient stimulus to the rise of algae. With this argument, an increase in P availability would have fueled larger predatory algae and generated fundamental shifts in the structure and importance of eukaryote ecosystems (28).
Overall, the evolution of eukaryotic ecosystems from the earliest recognition of eukaryote fossils in the Late Paleoproterozoic era through the apparent rise of algae in the Neoproterozoic era is very unclear. We can summarize these uncertainties with several key questions: 1) When did marine ecosystems acquire phagocytosis and eukaryotic photosynthesis? 2) How might nutrient limitation have impacted the importance of eukaryotes as ecosystem members? 3) How important was eukaryote photosynthesis in carbon cycling before the apparent rise of algae? 4) Why is the Late Paleoproterozoic and Mesoproterozoic era biomarker record so silent on the presence of early eukaryotes when the fossil record indicates they were there?
We explore these questions with a size- and trait-based model of unicellular plankton communities. Our modeling combines simple physics (e.g., diffusion, light acquisition), vital rates (e.g., feeding dynamics and efficiencies, growth rates and death rates), and energetic arguments (e.g., photosynthetic and growth efficiencies) with basic ecosystem properties (e.g., nutrient availability, viral attack, DOC availability) to make predictions about the activity and size spectrum of the different basic metabolic lifestyles. As the model is widely based on first principles (39), it has the generality to be applicable to exploring ancient organisms and their ecosystem dynamics.
Organism Size
Our modeling is linked to the geologic record through the size of microfossils preserved in the ancient record. Most of these microfossils are preserved as single-celled organisms, so-called acritarchs, defined as organic-walled vesicles with a central cavity and with an uncertain biological affinity. Therefore, acritarchs do not include obvious filamentous prokaryotic organisms.
While acritarchs have traditionally been viewed as the remains (both resting stages and vegetative states) of planktonic organisms (e.g., ref. 40), many of them may have been benthic as well (41). A benthic attribution for some morphospecies is based on their predominance in very shallow-water environments, where a benthic lifestyle seems most likely. Also, some of these shallow-water morphospecies are large, suggesting that they grew in a nutrient-rich environment as can be found in sediments (41). However, many other acritarch morphospecies are most likely the remains of planktonic organisms based on their broad geographical distribution (41). Conspicuous examples include various members of the sphaeromorphic genus Leiosphaeridia, which can be found in settings ranging from shallow to deep water and which are preserved in sedimentary rocks ranging from sandstones to shales (e.g., refs. 17, 42, and 43). Leiosphaeridia are viewed as possible eukaryotes (1, 17), and even though they were planktonic, they may also be quite large, with sizes ranging up to 300 to 400 μm for some specimens and median sizes of typically 50 μm for individual morphospecies (17, 42).
As many acritarchs are likely the resting stages, or cysts, of vegetative cells, one might ask how these cells compare in size with vegetative cells. Indeed, the resting stages of single-celled algae approximate the size of the vegetative cells (44), while for ciliates, cyst formation usually leads to a reduction in cell size and volume; indeed, ciliate cyst diameters typically range from 25 to 60% of those for vegetative cells (45). For parasitic flagellates, cyst stages also trend smaller than vegetative stages (46). Thus, the microfossils representing cysts in the geologic should not overrepresent the size of the vegetative cells that produced the cysts.
A compilation of cell size for Proterozoic-aged acritarchs (from the data in ref. 47) is shown in Fig. 1. Many of these cells are nonspherical, but their sizes have been corrected to nominal spherical volumes with the dimensions given in ref. 47. These data represent a compilation from literature reports and are not a comprehensive view of the size distribution of acritarchs as preserved in rocks; such data have not been reported, except for some morphospecies in a few environments (e.g., refs. 17 and 42). Also, there may be taphonomic biases favoring the preservation of large cells over small cells, another reason why the distribution of cell sizes in the ancient marine environment could have differed from the distribution reported here. Important for our arguments, however, is that from the Late Paleoproterozoic era through the Neoproterozoic era, acritarchs preserved in the geologic record can be quite large: typically in the microplankton range (>20 μm) with sizes up to 400 μm or more (Fig. 1).
Fig. 1.
We focus on the biomass distribution of cell sizes because it is an emergent ecosystem property depending on the energy (nutrients) fed into the system and the physiological capacities of the cells (e.g., whether they can consume smaller cells by phagotrophy or not). The size distribution of phototrophs, for example, responds to nutrient availability where as noted above, small phototrophs are favored under nutrient limiting conditions. Under nutrient limitation, smaller cells, with their larger surface area to volume ratios, have an advantage in terms of nutrient uptake and acquisition. Thus, picoplankton (<2 μm) and especially, single-celled cyanobacteria dominate in nutrient-starved gyre regions of the oceans, whereas the larger nanoplankton (2 to 20 μm) and microplankton (>20 μm) dominate when nutrients are more available (e.g., refs. 36 and 48–51).
More generally, one can describe relationships between cell size and trophic strategy from first-principle considerations of trade-offs between cost and benefit for different life strategies, where the dominant strategy will provide the highest acquisition rate of carbon and nutrients (e.g., refs. 52–55). The smallest organisms are osmotrophic prokaryotes obtaining most of their carbon and nutrients from diffusion across the cell wall. As size increases, the photon flux becomes sufficient to drive carbon fixation rates in excess of carbon acquisition rates by diffusion. Thus, photosynthesis becomes the dominant life strategy (54, 55). In even larger cells, the diffusion of inorganic nutrients begins to limit cell growth, and while these larger cells still can benefit from photosynthesis, they supplement their carbon and nutrient requirements through phagotrophy. These larger cells are dominated by mixotrophs (conducting both phototrophy and phagotrophy). The largest cells become inefficient at acquiring carbon through both photosynthesis and diffusion but quite efficient at “grazing” organisms from the water column and ingesting them by phagotrophy. Thus, the largest single-celled organisms are predominantly heterotrophic phagotrophs.
These general principles explain the size distribution of preferred life strategies in the oceans (52, 54, 55), but the actual transitions between the different dominant lifestyles depend on concentrations of dissolved labile organic carbon, nutrients, and light intensity (54). Indeed, as already noted, the size of dominant phototrophs in a marine ecosystem depends critically on nutrient availability. Overall, however, if the size distribution of organisms is an emergent property of the metabolic lifestyle of ecosystem members, then we can make predictions about the relative significance of different life strategies in the past with some knowledge of the size of ancient ecosystem members as introduced above.
Our size- and trait-based modeling approach allows us to examine the impact of nutrient limitation on the size distribution of organisms in a number of differed modeled ecosystems. We will compare the results of this model with the geologic record to evaluate the importance of eukaryotes in Late Paleoproterozoic through Early Neoproterozoic era marine ecosystems.
Size-Based Ecosystem Model
A growing number of models utilize size-structured approaches to explore marine ecology (e.g., refs. 52 and 56–58). These models do not resolve individual species, but they utilize general organism traits (inorganic nutrient uptake rates, nutrient affinities, cellular carbon quotas, etc.) as a function of size to define how individual populations [phototrophs and heterotrophs of various size, for example (59, 60)] respond to various degrees of nutrient and light limitation, typically in the framework of an ocean circulation model. Our modeling follows the approach outlined in refs. 39, 53–55, and 57. Like other size-based approaches, our model is conceptually simple as it does not consider individual species but rather, considers general traits of organisms as a function of size. A more detailed description of the model and our modeling approach is found in SI Appendix and in ref. 39.
Overall, our size-based model requires a rather small parameter set determined by a combination of first principles (diffusion of nutrients, cell geometry, fluid mechanics, and quantum efficiency for light affinity) or extracted from meta-analyses of size-based relationships (clearance rate and prey size preference for phagotrophy and maximum growth rate) (SI Appendix, Table S2) (39). The model combines these principles with a simple description of cell metabolism and the rule that bigger cells prey on smaller cells to determine the biomass size structure of the entire community. The model includes a large size range of unicellular plankton from small prokaryotes to large eukaryotic phagotrophs and mixotrophs but without multicellular eukaryotes. Individual cells are characterized by size, measured in units of carbon. It is assumed that all cells are generalists and that they have the potential to live by combinations of photosynthesis, osmotrophy (take up nutrients and DOC) and phagotrophy (grazing on other organisms). These combinations include mixotrophy, a combination of photosynthesis and phagotrophy. The actual trophic strategy (the ratio between osmo-/photo-/phagotrophy) of cells of a given size is an emergent property of the model and depends on the conditions introduced (nutrients and light).
The size- and trait-based model simulates the upper water column as a chemostat driven by light and vertical mixing of nutrients from a (unmodeled) deep layer (61). Nutrient input rates are controlled both by the mixing rate and by the nutrient concentration of the water mixed into the upper layer. There is also mixing out of the upper water column (chemostat) of nutrients, labile dissolved carbon, and cells. We accommodate marine geographic variability by exploring a range of deep nutrient concentrations and mixing rates. We also explore low nutrient concentrations to simulate levels possibly relevant to earlier in Earth history.
The model generates a size distribution of biomass that we group into categories of picoplankton (diameter of <2 μm), nanoplankton (2 to 20 μm), and microplankton (>20 μm). Fluxes of nutrients and organic carbon into or out of the cell are described as biomass-specific gains or losses. Gains describe the uptake of nutrients, DOC, food consumption by phagotrophy, and the carbon produced by photosynthesis (light harvesting), while losses include losses to respiration, grazing, and cell death due to viral infections. The DOC fueling osmotrophy in our modeling is produced through cell lysis by viruses and through smaller losses during phototrophy, unassimilated food during phagotrophy, and passive losses across the cell membrane.
We explored mixing rates ranging between 0.01 and 0.1 d−1. While our model approximates a chemostat, these mixing rates do not correspond to growth rates (as in a true chemostat) due to nutrient recycling in our modeled ecosystem. Our mixing rates are also difficult to compare directly with rates of upwelling or vertical exchange as estimated in the modern oceans. However, our choice of mixing rates, when combined with the range of P concentrations viewed as appropriate for the chemocline of the modern oceans (below and in SI Appendix), generates growth rates up to 0.2 to 0.3 d−1 (results are below), a good estimate for average maximum phytoplankton growth rates in the oceans (62). In this way, we couple mixing in our model with modern ocean dynamics. We further discuss our choice of mixing rates in SI Appendix.
We model phosphate as the limiting nutrient. The phosphate concentrations in our model reflect those mixed and advected into the upper mixed layer of the oceans and not the maximum deep-water values that currently range from ca. 1.5 μM in the north Atlantic Ocean to 3 μM in the north Pacific Ocean (63). We estimate the concentrations of phosphate mixed into the upper ocean from literature assessments of the nitrate flux into the photic zone (64) as outlined in the SI Appendix. Overall, for modern coastal areas representing the type of environment preserving most of the fossil record, a reasonable range of P concentrations mixed into the photic zone is between about 0.06 and 0.6 μM (SI Appendix). Considering this range, we explore P concentrations from 1 to 0.001 μM but highlight values of 0.6, 0.06, and 0.006 μM. The higher two concentrations are viewed as relevant for modern coastal oceans, while the lower values would be relevant for oceans with 10 times (or more) lower nutrient levels than today. Furthermore, we have fixed the photic zone depth to 50 m, but changing this depth will only change overall rates of primary production and not the patterns of cell size and trait distributions. We have chosen a light level of 30 E/m2 per second, which is sufficient to ensure that plankton are mainly limited by nutrients.
In what follows, we present four different modeling scenarios aimed at testing different hypotheses regarding the evolutionary history of eukaryotes and their involvement in Late Paleoproterozoic through Early Neoproterozoic era marine ecosystems. Each model scenario consists of 11,000 model simulations where a subset of the model input parameters was randomly assigned (within ±2σ) to a lognormal distribution. The randomized parameters are used to calculate the specific clearance rate, nutrient affinity, diffusive uptake, light harvesting, maximum growth rate, passive losses, metabolic rate, and mixing rate (within the range of 0.01 to 0.1 d−1). A range in these parameter values reflects natural variability between species, and sampling around a lognormal distribution ensures that model results reflect this variability. Our approach, therefore, gives a range of possible solutions to each model scenario, providing robustness to the model result. The mean and SD for each of the randomized parameters can be found in SI Appendix, Table S2.
Case 1: Picoplankton Phototrophy/No Grazing.
This case assumes that only cyanobacterial phototrophs and osmotrophs were present in the ecosystem. This model is consistent with an ecosystem before the evolution of eukaryotic phagotrophy and algal photosynthesis but with the possibility of eukaryotic osmotrophy. Thus, this case is compatible with a pre-LECA evolutionary scenario, where the ecosystem had stem-group osmotrophic eukaryotes but lacked crown-group eukaryotes capable of phagotrophy and photosynthesis. This case is compatible with the LECA late evolutionary scenario when applied to Proterozoic eon ecosystems. To simulate this scenario, we model an ecosystem where photosynthesis is restricted to the picoplankton, a size range where pelagic cyanobacterial photosynthesis is dominant (65). Furthermore, we do not allow grazing by phagotrophy but do allow osmotrophy through the whole size distribution of organisms.
Model results show that for all levels of phosphate input, biomass is completely restricted to the picoplankton (Figs. 2 and 3 and Table 1). There is, however, an expansion to larger picoplankton as P levels increase (Fig. 2). Also, throughout the range of P concentrations, carbon acquisition is dominated by DOC uptake (osmotrophy) over photosynthesis by 2/1 to 3/1 (Table 2). The distribution of gains and losses as a function of size for the three P levels of 0.6, 0.06, and 0.006 μM is shown in Fig. 4. Consistent with model assumptions, a shift to cells outside the picoplankton range can only happen through osmotrophy; however, the model shows that osmotrophy is unable to support larger cells. Losses occur by viral lysis (and respiration at a constant rate; not shown) that closely follows growth rate.
Fig. 2.
Fig. 3.
Fig. 4.
Table 1.
| Total biomass, µg C/L (% total) | |||
|---|---|---|---|
| Picoplankton (<2 µm) | Nanoplankton (2–20 µm) | Microplankton (20–200 µm) | |
| Case 1: Picoplankton phototrophy/no grazing | |||
| P = 0.006 µM | 5.71 (100) | 0.00 (0) | 0.00 (0) |
| P = 0.06 µM | 41.7 (100) | 0.00 (0) | 0.00 (0) |
| P = 0.6 µM | 174 (100) | 0.00 (0) | 0.00 (0) |
| Case 2: Full phototrophy/no grazing | |||
| P = 0.006 µM | 5.97 (99.7) | 0.02 (0.3) | 0.00 (0) |
| P = 0.06 µM | 43.6 (97.2) | 1.25 (2.8) | 0.00 (0) |
| P = 0.6 µM | 193 (74.9) | 60.3 (23.4) | 4.47 (1.7) |
| Case 3: Picoplankton phototrophy/grazing | |||
| P = 0.006 µM | 5.11 (90.4) | 0.54 (9.6) | 0.00 (0) |
| P = 0.06 µM | 27.9 (75.2) | 8.98 (24.2) | 0.23 (0.6) |
| P = 0.6 µM | 68.1 (69.0) | 28.9 (29.3) | 1.66 (1.7) |
| Case 4: Full phototrophy/grazing | |||
| P = 0.006 µM | 4.17 (72.4) | 1.52 (26.4) | 0.07 (1.2) |
| P = 0.06 µM | 21.7 (48.0) | 20.5 (45.3) | 3.02 (6.7) |
| P = 0.6 µM | 129 (49.4) | 102 (39.1) | 29.9 (11.5) |
Table 2.
| Carbon uptake, % of total | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Picoplankton (<2 μm) | Nanoplankton (2–20 μm) | Microplankton (20–200 μm) | |||||||
| Light | DOC | Phago | Light | DOC | Phago | Light | DOC | Phago | |
| Case 1: Picoplankton phototrophy/no grazing | |||||||||
| P = 0.006 µM | 36.6 | 63.4 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 |
| P = 0.06 µM | 33.2 | 66.8 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 |
| P = 0.6 µM | 26.0 | 74.0 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 |
| Case 2: Full phototrophy/no grazing | |||||||||
| P = 0.006 µM | 38.2 | 61.7 | 0.00 | 0.20 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 |
| P = 0.06 µM | 34.4 | 64.2 | 0.00 | 1.20 | 0.20 | 0.00 | 0.00 | 0.00 | 0.00 |
| P = 0.6 µM | 21.0 | 69.9 | 0.00 | 7.00 | 1.60 | 0.00 | 0.51 | 0.00 | 0.00 |
| Case 3: Picoplankton phototrophy/grazing | |||||||||
| P = 0.006 µM | 35.7 | 58.1 | 1.30 | 0.00 | 0.30 | 4.50 | 0.00 | 0.00 | 0.11 |
| P = 0.06 µM | 31.3 | 54.8 | 1.30 | 0.00 | 0.80 | 11.1 | 0.00 | 0.00 | 0.74 |
| P = 0.6 µM | 27.2 | 58.4 | 1.20 | 0.00 | 0.90 | 11.4 | 0.00 | 0.00 | 0.81 |
| Case 4: Full phototrophy/grazing | |||||||||
| P = 0.006 µM | 29.9 | 50.5 | 5.00 | 10.2 | 1.60 | 6.30 | 0.31 | 0.00 | 0.60 |
| P = 0.06 µM | 18.8 | 43.2 | 0.90 | 17.2 | 3.30 | 12.3 | 1.27 | 0.00 | 3.09 |
| P = 0.6 µM | 14.6 | 53.1 | 0.80 | 16.1 | 2.00 | 7.60 | 1.60 | 0.00 | 4.21 |
Light indicates carbon acquisition due to photosynthesis (photosynthesis). DOC indicates carbon acquisitions due to DOC uptake (osmotrophy). Phago indicates carbon uptake due to phagocystis (grazing).
Case 2: Full Phototrophy/No Grazing.
As discussed above, there are uncertainties as to when phagotrophy evolved as a eukaryotic feeding strategy. Therefore, we modeled a case without grazing (and thus, also without mixotrophy) but with phototrophy throughout the size range. This model is compatible with an evolutionary scenario where grazing by phagotrophy emerged after the evolution of algal photosynthesis. This is not a well-accepted evolutionary scenario, but it could be relevant if phagotrophy became efficient and ecologically important much latter in time than the initial phagotrophic acquisition of the cyanobacteria that become the chloroplast or alternatively, if the cyanobacterial symbiont was acquired by a process other than phagotrophy.
In this case, most of the biomass is distributed within the picoplankton but with a small tail of biomass distributed into the nanoplankton range as nutrient levels surpass about 0.06 μM (Figs. 2 and 3 and Table 1). In addition, a very small amount of microplankton emerge at the highest nutrient levels explored (Fig. 3 and Table 1). Among the picoplankton and at all nutrient levels, carbon acquisition is dominated by DOC uptake. In contrast, for those nutrient levels supporting a nanoplankton population, photosynthesis dominates carbon acquisition. At high nutrient levels (0.6 μM P), nanoplankton photosynthesis contributes to 25% of the total photosynthesis, while microplankton contribute to 1.8% of the total (Table 2). At intermediate nutrient levels (0.06 μM P), nanoplankton contribute to 3.4% of the total photosynthesis, with no contribution from the microplankton (Table 2). These trends in biomass distribution and carbon acquisition strategies are mirrored in the carbon losses and gains (Fig. 4).
Case 3: Picoplankton Phototrophy/Grazing.
This case simulates the evolutionary scenario where mitochondria were acquired early during eukaryogenesis. In this scenario, phagocytosis was a stem-group eukaryotic trait, and algal photosynthesis emerged after the phagocytosis of a cyanobacterium, the formation of the chloroplast, and the beginnings of algal photosynthesis. Thus, this case considers the time in between the evolution of phagocytosis and algal photosynthesis, and it is relevant for evolutionary models where Middle Proterozoic oceans had eukaryotic phagotrophy and osmotrophy but no algal photosynthesis. Instead, all photosynthesis was by cyanobacteria. We model this case by allowing photosynthesis only among the picoplankton, where cyanobacterial photosynthesis dominates, and with phagocytosis (grazing) and osmotrophy allowed throughout the size spectrum.
In this model, the distribution of biomass is very nutrient dependent. At low nutrient levels (0.006 μM P), about 90% of the biomass is distributed among the picoplankton, with the rest in the nanoplankton (Figs. 2 and 3 and Table 1). As nutrient levels increase to 0.06 μM, nanoplankton contribute to 24% of the total biomass, and this contribution increases to 29% at 0.6 μM P. With nutrient levels of 0.06 μM P and greater, microplankton contribute between 0.6 and 1.7% of the total biomass (Table 1). By definition in this case, carbon acquisition by photosynthesis is restricted to the picoplankton. Furthermore, phagotrophy is responsible for most of the carbon acquisition in the nanoplankton (92 to 93%) and all of the carbon acquisition in the microplankton. Regardless of nutrient levels, osmotrophy is only important among the picoplankton (Table 2). Carbon gains through photosynthesis follow model assumptions, but the addition of phagocytosis generates gains and losses in the larger sizes classes, where the gains by feeding are distributed to larger sizes than losses by grazing (Fig. 4). This difference in distribution reflects the size selectivity of phagotrophic grazers. With the incorporation of grazing, viral lysis becomes less important.
Case 4: Full Phototrophy/Grazing.
In this case, we allow for the full ecosystem of phototrophy, phagotrophy, and mixotrophy through the whole size range, although the model is restricted to single-celled organisms. This case simulates the evolutionary scenario where all single-celled eukaryotic feeding strategies have evolved and were part of the marine ecosystem from the Late Paleoproterozoic era and onward.
Model results show that as nutrient levels increase, biomass becomes more distributed into the larger size classes (Figs. 2 and 3 and Table 1). Indeed, at P levels of 0.006, 0.06, and 0.6 μM, nanoplankton account for 26, 45, and 39%, respectively, of the biomass, and microplankton biomass increases from 1.2 to 6.6 to 11.5% (Table 1). At all nutrient levels, nanoplankton photosynthesis contributes substantially to total photosynthesis, ranging from 25% at 0.006 μM P to 46% at 0.06 μM P and 50% at 0.6 μM P. Through the same nutrient increases, microplankton photosynthesis increases from 1.2 to 6.6%, remaining at 6.6% of the total photosynthesis at the highest nutrient level explored. Phagotrophy is an important mode of carbon acquisition at all nutrient levels, but it becomes relatively more important among the microplankton as nutrient levels increase, and it is the most important mode of microplankton carbon acquisition at all nutrient concentrations (Table 2). With the full ecosystem model, carbon gains by photosynthesis are distributed to large cell sizes, and this encourages elevated growth among the nanoplankton compared with case 3 (Fig. 4). Like in case 3, carbon gains by phagocytosis are distributed to larger sizes than losses by grazing.
Discussion
General Considerations.
Our modeling explores how differentially adding organismal traits (osmotrophy, phototrophy, and phagotrophy) changes the size distribution of organisms and their metabolic function in a marine ecosystem. We evaluate these results considering the fossil record of organism size and a variety of scenarios for eukaryote evolution. The model’s strength is that its assumptions are universal, either based on diffusion physics and conservation of mass or based on general allometric relations to cell size, including nutrient affinities, light affinities, feeding rates, and growth rates, with additional constraints on prey size preference, metabolic rates, and rates of viral infection (SI Appendix). These relations are not tied to specific species of plankton and can plausibly be taken as the limits within which evolution can operate, now and in the past. Ecosystem structure—the size structure and productivity of the plankton community—emerges as a product of these universal relations and the environmental conditions.
Key environmental variables in our modeling include nutrient levels, mixing rates into our modeled ecosystem, the depth of the photic zone, and light intensity. Considering these variables, we have simulated ecosystems across several orders of magnitude of phosphorus concentrations, over a wide range of values around the mean of the allometric relationships (SI Appendix), and with an order of magnitude range of mixing rates that reflect the rates experienced in modern coastal environments (SI Appendix).
The model’s conceptual simplicity does not reflect the true dynamics of the marine environment, including physical circulation, organism migration, and particle settling. Still, the mixing into our model approximates mixing dynamics in the oceans, and particle settling is approximated by mixing out of the modeled environment. Overall, we believe that the general patterns of ecosystem structure as generated by our model are robust predictions.
In what follows, we start with general discussions of organism size and continue by addressing the four questions raised in the Introduction.
Considerations of Size.
Size enters in our discussion in two important ways. First, we must consider whether the cell sizes reproduced in our modeled ecosystem can be best ascribed to eukaryotes or prokaryote affinity. Taking clues from pelagic organisms in the modern oceans, prokaryotes typically have maximum cell volumes in the 0.5- to 0.6-μm3 range (66–68), yielding nominal (assuming a sphere) maximum cell diameters of about just under 1 μm. Also, the dominant pelagic cyanobacterial genera, Prochlorococcus and Synecococcus, have typical diameters of 0.6 and 0.9 μm, respectively (69). Furthermore, marine nanoplankton and microplankton are dominated by eukaryotes (65). Thus, with modern pelagic marine ecosystems as a guide, prokaryotes should be mostly found among the picoplankton, and eukaryotes should be mostly found among the nanoplankton and microplankton (and larger).
There are, however, notable exceptions. Picoplankton-sized eukaryotes exist, and in the modern ocean, phototrophic picoeukaryotes (<3 μm) can comprise from about 11 to 70% (by mass) of the picoplankton-sized phototrophic community (65, 70) and 3 to 25% of the total picoplankton community, including prokaryotic heterotrophs (65). The existence of picoeukaryotes, however, does not impact any of our conclusions as our focus is on large organisms and particularly, those in the microplankton range, where many fossil acritarchs are found (Fig. 1).
Large prokaryotes, however, also exist. Thus, pelagic cyanobacteria can reach large sizes where in a conspicuous example, filaments of the bundle-forming, nitrogen-fixing Trichodesmium sp. can reach widths of up 10 μm and lengths of up to 40 μm (71). A range of benthic prokaryotes, including a variety of cyanobacteria and sulfide-oxidizing bacteria, can also attain dimensions well into the microplankton size range (72). However, benthic prokaryotes are not relevant for the modeling conducted here. Furthermore, large pelagic filamentous cyanobacteria are dominantly nitrogen-fixing organisms. The most important of these in the modern oceans is Trichodesmium sp. While globally important in nitrogen fixation (e.g., refs. 73–75) and locally important in carbon fixation in some cases (e.g., ref. 76), Trichodesmium sp. are not major global contributors to primary production, accounting for an estimated 0.6 to 1% of global rates of primary production (77). Thus, large filamentous cyanobacteria do not typically dominate marine pelagic prokaryote populations.
Next, we consider more specifically whether the size of microfossils preserved in the geologic record allows general statements about their phylogenetic affinity. Given the prospect of large prokaryotes, particularly in benthic environments, size alone may be a poor indicator of phylogenetic affinity when considering the fossil record of microfossils. This point has been made in the literature (e.g., ref. 29). However, as discussed above, many large microfossils preserved in the fossil record contain a number of characteristics indicating eukaryote affinity. So, even if large microfossils preserved in the geologic record were not all eukaryotes, large eukaryotes were certainly present in ancient Late Paleoproterozoic through Early Neoproterozoic era marine pelagic ecosystems.
We now consider the questions raised in the Introduction.
When Did Marine Ecosystems Acquire Phagocytosis and Eukaryotic Photosynthesis?
As noted above, the earliest eukaryotic microfossils date to about 1,700 Ma. Many of these eukaryotic fossils are quite large, with sizes concentrated in the microplankton range, although some range up to 400 μm, well into the mesoplankton range (200 to 2,000 μm) (Fig. 1). These fossils, however, do not preserve features allowing definitive placement into specific eukaryotic clades, so their metabolic lifestyles are unknown. Our size-based modeling, however, can help define the metabolic nature of eukaryotes in the ecosystems where these ancient organisms lived.
Thus, from our modeling, the large sizes of microfossils preserved in the Late Paleoproterozoic through Neoproterozoic geologic record (Fig. 1) are incompatible with ecosystems lacking phagocytosis. Thus, in case 1, an evolutionary scenario with osmotrophy and picoplankton photosynthesis but no phagotrophy, biomass is restricted to the picoplankton size range (Fig. 3 and Table 1). If we add the possibility of photosynthesis throughout the size range, as in case 2, there is an increase in cell mass in the nanoplankton, especially at higher nutrient levels. However, even at the highest nutrient concentrations, microplankton account for less than 2% of the total biomass (Figs. 2 and 3 and Table 1), and these would be phototrophs (Table 2). Case 2 could satisfy the geologic record if all acritarchs in the Late Paleoproterozoic through Early Neoproterozoic era marine ecosystems were phototrophs but minor ecosystem members, growing under high nutrient levels. There is no agreement, however, that all microplankton-sized Proterozoic-aged acritarchs were phototrophs, and furthermore, case 2 is built on the unlikely evolutionary scenario that efficient eukaryotic phototrophy preceded phagotrophy.
The more likely possibility is that ecosystems lacking phagocytosis do not satisfy the geologic record. Our modeling also indicates that large acritarchs of the geologic record were likely not osmotrophs. Parameter choices dictate to some extent the size range of osmotrophic organisms, but first-principle considerations indicate that osmotrophy is the most efficient carbon acquisition strategy only for small organisms (39, 53).
In contrast, models with phagotrophy produce biomass distributions well into the size range of eukaryotes preserved in the microfossil record (Fig. 2 and Table 1). There are, however, differences in the model scenarios. In case 3, where photosynthesis is restricted to the picoplankton but grazing by phagotrophy is allowed throughout the size range, distinct populations of picoplankton and nanoplankton are produced, with only small amounts of microplankton accounting for 1.7% of the total biomass in the highest nutrient scenario (Figs. 2 and 3 and Table 1). These microplankton are phagotrophs (Table 2). Case 4 (with osmotrophy and photosynthesis throughout the size range and with grazing by phagocytosis) produces organisms well in the microplankton size range, with microplankton biomass ranging from 1.2% of the total biomass at 0.006 μM P to 6.6 and 11.5% at 0.06 and 0.6 μM P, respectively (Table 2). These organisms gain carbon through a mix of phagotrophy and photosynthesis, where mixotrophy is likely an important lifestyle. Thus, both cases 3 and 4 might be consistent with the geologic record, especially at high nutrient levels, but with different predictions about the abundance of eukaryotes in the ecosystem and especially, the presence or absence of eukaryotic phototrophy.
We appeal to the fossil record to differentiate between these cases. Thus, apparent crown-group red algal fossils were found in the 1,600-Ma Chitakoot Formation of the Vindhyan basin, India (78), while large decimeter-sized carbonaceous compressions from the 1,560-Ma Gaoyuzhuang Formation, northern China (4) appear to be multicellular eukaryotes and may have been photosynthetic. In addition, high-temperature pyrolysis of kerogens from the ca. 1,400-Ma Xiamaling Formation of north China released the full suite of C27 (cholestane), C28 (ergostane), and C29 (stigmastane) steranes (79). As in other studies, no steranes were found in bitumens extracted from the Xiamaling rocks, but the presence of C27, C28, and C29 in the kerogens demonstrates the presence of both red and green algae in the ecosystem at this time (79). Some Late Paleoproterozoic– and Mesoproterozoic-aged acritarchs have also been interpreted as algae (10). Therefore, the fossil record supports phototrophy as a eukaryotic trait as far back as 1,600 or before, making case 4, the full ecosystem model, the most likely representation of Late Paleoproterozoic through Early Neoproterozoic era marine ecosystems.
As a final note, our full ecosystem model (case 4) assumes that mixotrophy was possibly an important metabolic strategy. While mixotrophic fossils cannot be identified in the fossil record, Late Paleoproterozoic and Mesoproterozoic mixotrophy would be consistent with the timing of eukaryote evolution as informed by our modeling. Thus, through our modeling, we place the evolution of both phagocytosis and algal photosynthesis to at least 1,600 to 1,700 Ma, and as the incorporation of the original cyanobacterium forming the chloroplast was believed to occur though phagocytosis, the original algae were likely also mixotrophic. It is logical to assume that this lifestyle would have persisted through subsequent geologic time.
How Might Nutrient Limitation Have Impacted the Importance of Eukaryotes as Ecosystem Members?
We argue above that case 4, our full ecosystem model, is the most appropriate in describing Late Paleoproterozoic through Early Neoproterozoic era marine ecosystem dynamics. From this model, significant microplankton (>20 μm), as revealed in the fossil record (Fig. 1), require a minimum phosphate concentration of 0.06 μM mixed into the upper ocean (Fig. 2 and Table 1). Given the nature of our model, we cannot precisely relate these phosphate concentrations to those of the deep oceans, and as such, our model offers clues as to the limits of Proterozoic deep-water P levels but not precise estimates. Still, as noted above (SI Appendix), we argue that phosphate concentrations between 0.06 and 0.6 μM in our model reflect modern-day P concentrations. As noted in the SI Appendix, these P concentration estimates are not completely independent of choices for mixing rate. For example, a high mixing rate of 0.2 d−1 would correlate with a lower minimum estimate of P concentration for the modern ocean by a factor of two (SI Appendix). Higher mixing rates, however, are untenable (SI Appendix), and we believe that the range of mixing rates we have considered, 0.01 to 0.1 d−1, likely reflects the dynamics of modern coastal oceans. Thus, our range of modern chemocline P concentrations likely represent reasonable estimates.
Overall, our modeling suggests that a substantial microplankton population is compatible with deep-water phosphorus levels comparable with today. Lower levels of P are also possible, but deep-water P levels of 10 times lower than today (with chemocline values ranging from 0.006 to 0.06 μM P) would just barely supply enough P at the upper end of concentration estimates (0.06 μM) to produce a limited microplankton population (Fig. 2 and Table 1).
Our modeling results are largely compatible with those of ref. 38, which also explored the role of nutrient concentrations on ecosystem structure with a trait-based ecosystem model. Their model is based on a three-dimensional ocean model allowing ca. 1,000-km-scale resolution of plankton types. Most of their model results were obtained without mixotrophy, and carbon cycling by the different trophic types was not reported. Still, at bottom-water nutrient levels 100 times less than today, the model produced only picoplankton-sized phototrophs. At nutrient levels 10 times less than today, the model produced mainly picoplankton- and nanoplankton-sized phototrophs, with a very minor contribution of small microplankton in a few regions of the global ocean. Thus, in calibrating with the fossil record, the model results of ref. 38 would also suggest Late Paleoproterozoic through Early Neoproterozoic deep-water phosphate concentrations of between modern levels and 10 times less, a result not dissimilar to ours.
How Important Was Eukaryote Photosynthesis in Carbon Cycling before the Apparent Rise of Algae?
Our model results challenge the view that eukaryotes were unimportant ecosystem members until the Early to Middle Neoproterozoic era in what is known as the rise of algae (e.g., refs. 28, 30, and 80). Indeed, if we accept that case 4 is the most appropriate and that chemocline P levels of 0.06 to 0.6 μM were most likely, then nanoplankton and larger organisms were equal in mass to the picoplankton (Table 1). If most of these organisms were eukaryotes, as would be probable based on their large size, then from the Late Paleoproterozoic through the Early Neoproterozoic eras, eukaryotes accounted for approximately one-half of the biomass in the oceans.
We further assess the importance of eukaryotes in photosynthesis. From case 4 and for chemocline P levels between 0.06 and 0.6 μM, organisms of nanoplankton and larger size account for ≥50% of total carbon production through photosynthesis (Table 2). As these organisms would have been dominantly eukaryotes based on their size, our modeling suggests an active role for eukaryotic photosynthesis in the marine carbon cycle from the Late Proterozoic era and onward. As noted above, large filamentous cyanobacteria are found. However, most of these are either benthic, or nitrogen fixers when pelagic. With the modern oceans as a guide, these pelagic nitrogen fixers are crucial in balancing the marine nitrogen cycle but very small contributors to marine photosynthesis, except locally. Thus, most nanoplankton and larger photosynthetic organisms from the Late Paleoproterozoic through Early Neoproterozoic would likely have been eukaryotic.
We emphasize that our model results derive from the basic requirements of organisms to maximize their growth as a function of size utilizing the basic ecosystem functions of photosynthesis, respiration, osmotrophy, grazing, viral lysis, and respiration. Thus, if ecosystem members were available to conduct these functions, simple considerations of maximizing energy gain dictate the distribution of functional types as a function of organism size. The model results are quite general.
Why Is the Late Paloeproterozoic and Mesoproterozoic Biomarker Record so Silent on the Presence of Early Eukaryotes When the Fossil Record Indicates They Were There?
Our results suggest that Late Paleoproterozoic through Early Neoproterozoic era marine ecosystems would have been rich in eukaryotic organisms, including algae, mixotrophs, and grazers. This scenario, while compatible with the Proterozoic era microfossil record as noted above, is not obviously compatible with the record of fossil steranes derived from bitumen liberated from ancient rocks. Thus, as also noted above, numerous studies, when carefully controlled for contamination, have failed to find sterane biomarkers in bitumen extracts of sedimentary rocks older than 780 Ma (28, 30, 81–87). Indeed, these biomarker results, combined with the appearance of bitumen-derived sterane biomarkers in rocks younger than 780 Ma, have given rise to the idea of the rise of algae in the Tonian period.
We can, however, pose the question differently and ask the following. Why does the biomarker record of extracted bitumen show no evidence for eukaryotes in rocks where a diverse array of large and likely eukaryotic microfossils is found? For example, sediments of the ca. 1,500-Ma Roper Group of Australia contain a diverse array of likely eukaryotic microfossils, including T. plana, V. lophostriata, and Dictosphaera sp. (88), but the biomarker record of extracted bitumen yields no steranes (81, 82). The same is true for the ca. 1,380-Ma Xiamaling Formation, where an abundant and diverse array of eukaryotic microfossils is found (16) in the absence of steranes from extracted bitumen (79, 83).
One possible answer is that eukaryotes in these ancient ecosystems were sufficiently abundant to leave a significant microfossil record but insufficiently abundant to leave a record of extractable steranes. This possibility is difficult to rule out. Proterozoic eon microfossil reports do not systematically explore microfossil abundance, and even if they did, it would be difficult to translate the abundance of microfossils in rocks into abundance as ecosystem members. However, a lack of quantifiable ecosystem abundance does mean that eukaryotes were minor ecosystem members, and this suggestion conflicts with our modeling results. Furthermore, fossil eukaryotes might have been more prevalent than the microfossil record indicates if there was taphonomic loss of smaller eukaryotes, potentially compounding the problem of an invisible bitumen sterane record.
Another suggestion is that any eukaryotes before the Neoproterozoic era lacked steranes, a flexible membrane, and that they were incapable of phagocytosis. If so, these stem-group eukaryotes did not have steranes as a preservable biomarker. This model is compatible with the suggestions of refs. 24 and 25 for the timing of major events in eukaryote evolution. However, as explored above, the large sizes of eukaryotes preserved in the fossil record require active grazing by phagotrophy and are thus incompatible with a Neoproterozoic acquisition of steranes in the eukaryote lineage.
A final suggestion is that steranes are less stable over time than other biomarker classes, like hopanes, where sterane and hopane concentrations are often compared, yielding an estimate of relative eukaryote to prokaryote abundance in ancient ecosystems (e.g., refs. 28 and 30). Indeed, when artificially heated, steranes were lost preferentially to hopanes in sedimentary rocks from the Permian Dalong Formation (89), while biodegradation also favors sterane loss relative to hopanes (90). Heating and biodegradation have not necessarily led to preferential sterane loss in Late Paleoproterozoic and Mesoproterozoic rocks, but the action of these processes on sterane preservation points to a differential stability between sterane and hopane biomarkers. In addition, as noted above, steranes have been found tightly bound to kerogen in several samples of the ca. 1,380-Ma Xiamaling Formation (79). Bitumen extracts of these rocks yielded no steranes, and the kerogen-bound steranes where only released after prolonged high-temperature pyrolysis, showing: 1) that steranes were indeed preserved in the Mesoproterozoic rocks of the Xiamaling Formation and 2) that the more lightly bound steranes typically released during bitumen extraction were somehow lost over time, leaving the steranes tightly bound in the kerogen.
Thus, we argue that the lack of steranes in bitumens from the pre–780-Ma biomarker is a problem of sterane preservation and not due to a lack of eukaryotes in Late Paleoproterozoic and Mesoproterozoic ecosystems. Relative sterane instability would explain the biomarker record and align our model results with the geologic record, implying that eukaryotes were major ecosystem members from 1,600 to 1,700 Ma and onward.
Anaerobic Protists?
In addition to answering the four questions raised in the Introduction, we also address the possibility that Late Paleoproterozoic– and Neoproterozoic-aged eukaryotes were anaerobes. If true, the fossil record of acritarchs preserves anaerobic eukaryotes. Indeed, there are speculations that aerobic eukaryotes may not have emerged until Neoproterozoic times and that earlier eukaryotes were anaerobic organisms (24, 25, 91). Our modeling is agnostic as to the oxygen levels where organisms lived, but we note that nanoplankton and microplankton eukaryotes can be found in anoxic water columns and sediments (e.g., refs. 92 and 93). Thus, one could postulate that ancient marine ecosystems from ca. 1,600 to 1,700 Ma, and forward to the rise of algae, were split into an upper oxygenated water column with cyanobacterial photosynthesis and a lower anoxic water column with anaerobic protists living by phagotrophy. In this scenario there is no algal photosynthesis and the anaerobic protists live from settling cyanobacterial bioamss. This scenario could allow for a late evolution of aerobic eukaryotes, including algae, and it would at least partially explain the lack of Paleo- and Mesoproterozoic algal steranes in the traditional bitumen biomarker record (although it would not explain the presence of kerogen-bound algal steranes as discussed above).
We have not modeled this case explicitly, but it most closely resembles our case 3 with photosynthesis limited to the picoplankton and with grazing by phagotrophy through the full size range. However, different from our model, in the absence of aerobic protists, the anaerobic phagocystis of cyanobacterial biomass would have been spatially and temporally detached from cyanobacterial production. Case 3 produces significant microplankton only at the highest P levels, and the spatial/temporal detachment of grazing from primary production would reduce the energy available to the grazers, likely reducing their abundance at a given level of P when compared with case 3 results. Thus, a scenario with the late emergence of aerobic protists would be less compatible with the geologic record of acritarch sizes than our case 3 results and would likely be only tenable at the highest levels of P. Furthermore, the late emergence of aerobic protists would be incompatible with evidence for photosynthetic algae by ca. 1600 Ma as discussed above, except under the very unlikely scenario that algal photosynthesis evolved long before the evolution of aerobic phagotrophic grazing. However, we see no obvious reason why aerobic protist phagotrophy would be absent if there was sufficient membrane flexibility to allow for anaerobic phagotrophy and if the aerobic metabolisms associated with algal photosynthesis had already evolved. Altogether, our model results, combined with the geologic record, are not compatible with the absence of aerobic protistan grazers from the time window of ca. 1600 Ma and forward.
Conclusions
It is generally thought that eukaryotes did not become major members of marine ecosystems until the rise of algae beginning around 780 Ma. Despite this, the microfossil record preserves eukaryotic fossils from at least 1,700 Ma. Many of these fossils are also quite large. We used a size- and trait-based ecosystem model to show that the large microfossils preserved in the geologic record are consistent with marine ecosystems housing abundant eukaryotic members, including osmotrophs, phototrophs, mixotrophs, and grazers. This conclusion is strengthened with independent evidence for phototrophy during the Mesoproterozoic era. From ca. 1,700 Ma onward, these eukaryotes likely contributed to 50% or more of the standing biomass and about 50% of marine primary production. Our modeling is consistent with deep-water phosphate concentrations similar to present, although they could have been as much as 10% of present levels. We suggest that the general lack of eukaryotic steranes in the pre–780-Ma fossil record is due to poor sterane preservation compared with other biomolecules, like hopanes.
Data, Materials, and Software Availability
Previously published data were used for this work (44). The model used can be found at: https://github.com/trinefrisbaek/NUMmodel_Proterozoic_eukaryotes (94).
Acknowledgments
We thank Karl Attard for his help with the graphical software and Shuhai Xiao for sharing his database on acritarch size distribution through the Proterozoic eon. We also acknowledge discussions with Susan Porter and Nick Butterfield. The work was supported by Villum Fonden Grant 16518 and by the Villum Kahn Rasmussen (VKR) Centre of Excellence “Ocean Life.”
Supporting Information
Appendix 01 (PDF)
- Download
- 1.17 MB
References
1
A. H. Knoll, E. J. Javaux, D. Hewitt, P. Cohen, Eukaryotic organisms in Proterozoic oceans. Philos. Trans. R. Soc. Lond. B Biol. Sci. 361, 1023–1038 (2006).
2
E. Javaux, “Early eukaryotes in Precambrian oceans” in Origins and Evolution of Life: An Astrobiological Perspective, M. Gargaud, P. López-Garcia, H. Martin, Eds. (Cambridge University Press, Cambridge, United Kingdom, 2011), chap. 26, pp. 414–449.
3
N. J. Butterfield, Early evolution of the eukaryota. Palaeontology 58, 5–17 (2015).
4
S. Zhu et al., Decimetre-scale multicellular eukaryotes from the 1.56-billion-year-old Gaoyuzhuang Formation in North China. Nat. Commun. 7, 11500 (2016).
5
Y. Z. Yan, S. X. Zhu, “Discovery of acanthomorphic acritarchs from the Baicaoping Formation in Yongji, Shanxi and its geological significance” (Tech. Rep., 1992). Acta Micropalaeontol. Sin. 9 (3), 267-282. In Chinese, with English abstract.
6
S. Xiao, A. H. Knoll, A. J. Kaufman, L. Yin, Y. Zhang, Neoproterozoic fossils in Mesoproterozoic rocks? Chemostratigraphic resolution of a biostratigraphic conundrum from the North China Platform. Precambrian Res. 84, 197–220 (1997).
7
E. J. Javaux, A. H. Knoll, M. R. Walter, TEM evidence for eukaryotic diversity in mid‐Proterozoic oceans. Geobiology 2, 121–132 (2004).
8
G. Hu, T. Zhao, Y. Zhou, Depositional age, provenance and tectonic setting of the Proterozoic Ruyang Group, southern margin of the North China Craton. Precambrian Res. 246, 296–318 (2014).
9
W. Su et al., Luoyu and Ruyang Group at the south margin of the North China Craton (NCC) should belong in the Mesoproterozoic Changchengian System: Direct constraints from the LA-MC-ICPMS U-Pb age of the tuffite in the Luoyukou Formation, Ruzhou, Henan, China. Geol. Survey Res. 35, 96–108 (2012).
10
M. Moczydłowska, E. Landing, W. Zang, T. Palacios, Proterozoic phytoplankton and timing of chlorophyte algae origins. Palaeontology 54, 721–733 (2011).
11
Z. R. Adam, M. L. Skidmore, D. W. Mogk, N. J. Butterfield, A Laurentian record of the earliest fossil eukaryotes. Geology 45, 387–390 (2017).
12
Z. Zhongying, Clastic facies microfossils from the Chuanlinggou formation (1800 Ma) near Jixian, North China. J. Micropalaeontol. 5, 9–16 (1986).
13
Y. Peng, H. Bao, X. Yuan, New morphological observations for Paleoproterozoic acritarchs from the Chuanlinggou Formation, North China. Precambrian Res. 168, 223–232 (2009).
14
K. Pang, Q. Tang, X.-L. Yuan, B. Wan, S. Xiao, A biomechanical analysis of the early eukaryotic fossil Valeria and new occurrence of organic-walled microfossils from the Paleo-Mesoproterozoic Ruyang Group. Palaeoworld 24, 251–262 (2015).
15
H. Agić, M. Moczydłowska, L. Yin, Diversity of organic-walled microfossils from the early Mesoproterozoic Ruyang Group, North China Craton–A window into the early eukaryote evolution. Precambrian Res. 297, 101–130 (2017).
16
L. Miao, M. Moczydłowska, M. Zhu, A diverse organic-walled microfossil assemblage from the Mesoproterozoic Xiamaling Formation, North China. Precambrian Res. 360, 106235 (2021).
17
E. J. Javaux, A. H. Knoll, Micropaleontology of the lower Mesoproterozoic Roper Group, Australia, and implications for early eukaryotic evolution. J. Paleontol. 91, 199–229 (2017).
18
A. H. Knoll, D. Lahr, “Fossils, feeding, and the evolution of complex multicellularity” in Multicellularity, Origins and Evolution, K. J. Niklas, S. A. Newman, Eds. (The Vienna Series in Theoretical Biology, MIT Press, Cambridge, MA, 2016), chap. 1, pp. 3–16.
19
P. Forterre, A new fusion hypothesis for the origin of Eukarya: Better than previous ones, but probably also wrong. Res. Microbiol. 162, 77–91 (2011).
20
T. Cavalier-Smith et al., Multigene eukaryote phylogeny reveals the likely protozoan ancestors of opisthokonts (animals, fungi, choanozoans) and Amoebozoa. Mol. Phylogenet. Evol. 81, 71–85 (2014).
21
A. M. Poole, S. Gribaldo, Eukaryotic origins: How and when was the mitochondrion acquired? Cold Spring Harb. Perspect. Biol. 6, a015990 (2014).
22
V. L. Koumandou et al., Molecular paleontology and complexity in the last eukaryotic common ancestor. Crit. Rev. Biochem. Mol. Biol. 48, 373–396 (2013).
23
W. F. Martin, A. G. M. Tielens, M. Mentel, S. G. Garg, S. B. Gould, The physiology of phagocytosis in the context of mitochondrial origin. Microbiol. Mol. Biol. Rev. 81, e00008–e00017 (2017).
24
D. B. Mills, The origin of phagocytosis in Earth history. Interface Focus 10, 20200019 (2020).
25
S. M. Porter, Insights into eukaryogenesis from the fossil record. Interface Focus 10, 20190105 (2020).
26
N. Yutin, M. Y. Wolf, Y. I. Wolf, E. V. Koonin, The origins of phagocytosis and eukaryogenesis. Biol. Direct 4, 9 (2009).
27
D. A. Baum, B. Baum, An inside-out origin for the eukaryotic cell. BMC Biol. 12, 76 (2014).
28
J. J. Brocks et al., The rise of algae in Cryogenian oceans and the emergence of animals. Nature 548, 578–581 (2017).
29
A. H. Knoll, Paleobiological perspectives on early eukaryotic evolution. Cold Spring Harb. Perspect. Biol. 6, 1–14 (2014).
30
J. A. Zumberge, D. Rocher, G. D. Love, Free and kerogen‐bound biomarkers from late Tonian sedimentary rocks record abundant eukaryotes in mid‐Neoproterozoic marine communities. Geobiology (2019).
31
J. F. Allen, B. Thake, W. F. Martin, Nitrogenase inhibition limited oxygenation of earth’s proterozoic atmosphere. Trends Plant Sci. 24, 1022–1031 (2019).
32
S. Ratti, A. H. Knoll, M. Giordano, Grazers and phytoplankton growth in the oceans: An experimental and evolutionary perspective. PLoS One 8, e77349 (2013).
33
S. Porter, The rise of predators. Geology 39, 607–608 (2011).
34
C. T. Reinhard et al., Evolution of the global phosphorus cycle. Nature (2016).
35
C. Jones, S. Nomosatryo, S. A. Crowe, C. J. Bjerrum, D. E. Canfield, Iron oxides, divalent cations, silica, and the early earth phosphorus crisis. Geology 43, 135–138 (2015).
36
S. W. Chisholm, “Phytoplankton size” in Primary Productivity and Biogeochemical Cycles in the Sea, P. G. Falkowski, A. D. Woodhead, Eds. (Springer, 1992), pp. 213–237.
37
L. K. Eckford-Soper, D. E. Canfield, The global explosion of eukaryotic algae: The potential role of phosphorus? PLoS One 15, e0234372 (2020).
38
C. T. Reinhard et al., The impact of marine nutrient abundance on early eukaryotic ecosystems. Geobiology 18, 139–151 (2020).
39
K. H. Andersen, A. W. Visser, From first principles to structure and function of unicellular plankton communities. bioRxiv [Preprint] (2022). 10.1101/2022.05.16.492092. Accessed 25 May 2022.
40
C. Mendelson, “Acritarchs and prasinophytes” in Fossil Prokaryotes and Protists, J. H. Lipps, Ed. (Blackwell, Boston, MA, 1993), chap. 6, pp. 77–104.
41
N. J. Butterfield, Plankton ecology and the Proterozoic-Phanerozoic transition. Paleobiology 23, 247–262 (1997).
42
H. J. Hofmann, G. Jackson, Shale-facies microfossils from the Proterozoic Bylot Supergroup, Baffin Island, Canada. Paleontological Society Memoir 37, 1–39 (1994).
43
Z. R. Adam, M. L. Skidmore, D. W. Mogk, “Paleoenvironmental implications of an expanded microfossil assemblage from the Chamberlain Formation, Belt Supergroup, Montana” (Tech. Rep.) Geological Society of America Special Papers, 101–119. 10.1130/2016.2522(04).
44
M. Ellegaard, S. Ribeiro, The long-term persistence of phytoplankton resting stages in aquatic ‘seed banks.’ Biol. Rev. Camb. Philos. Soc. 93, 166–183 (2018).
45
F. Verni, G. Rosati, Resting cysts: A survival strategy in Protozoa Ciliophora. Ital. J. Zool. (Modena) 78, 134–145 (2011).
46
A. L. Leber, S. Novak-Weekley, “Intestinal and urogenital amebae, flagellates, and ciliates” in Manual of Clinical Microbiology, J. Versalovic et al., Eds. (American Society of Microbiology, ed. 10, 2011), pp. 2149–2171.
47
J. W. Huntley, S. H. Xiao, M. Kowalewski, 1.3 Billion years of acritarch history: An empirical morphospace approach. Precambrian Res. 144, 52–68 (2006).
48
A. J. Irwin, Z. V. Finkel, O. M. Schofield, P. G. Falkowski, Scaling-up from nutrient physiology to the size-structure of phytoplankton communities. J. Plankton Res. 28, 459–471 (2006).
49
E. Marañón, Cell size as a key determinant of phytoplankton metabolism and community structure. Annu. Rev. Mar. Sci. 7, 241–264 (2015).
50
M. Schartau, M. R. Landry, R. A. Armstrong, Density estimation of plankton size spectra: A reanalysis of IronEx II data. J. Plankton Res. 32, 1167–1184 (2010).
51
R. J. Brewin et al., Influence of light in the mixed-layer on the parameters of a three-component model of phytoplankton size class. Remote Sens. Environ. 168, 437–450 (2015).
52
B. A. Ward, S. Dutkiewicz, O. Jahn, M. J. Follows, A size‐structured food‐web model for the global ocean. Limnol. Oceanogr. 57, 1877–1891 (2012).
53
K. H. Andersen et al., Characteristic sizes of life in the oceans, from bacteria to whales. Annu. Rev. Mar. Sci. 8, 217–241 (2016).
54
S. Chakraborty, L. T. Nielsen, K. H. Andersen, Trophic strategies of unicellular plankton. Am. Nat. 189, E77–E90 (2017).
55
K. H. Andersen, D. L. Aksnes, T. Berge, Ø. Fiksen, A. Visser, Modelling emergent trophic strategies in plankton. J. Plankton Res. 37, 862–868 (2015).
56
B. A. Ward, S. Dutkiewicz, M. J. Follows, Modelling spatial and temporal patterns in size-structured marine plankton communities: Top-down and bottom–up controls. J. Plankton Res. 36, 31–47 (2014).
57
C. Serra-Pompei, F. Soudijn, A. W. Visser, T. Kiørboe, K. H. Andersen, A general size-and trait-based model of plankton communities. Prog. Oceanogr. 189, 102473 (2020).
58
R. F. Heneghan et al., A functional size-spectrum model of the global marine ecosystem that resolves zooplankton composition. Ecol. Modell. 435, 109265 (2020).
59
E. Litchman, C. A. Klausmeier, O. M. Schofield, P. G. Falkowski, The role of functional traits and trade-offs in structuring phytoplankton communities: Scaling from cellular to ecosystem level. Ecol. Lett. 10, 1170–1181 (2007).
60
N. S. Banas, Adding complex trophic interactions to a size-spectral plankton model: Emergent diversity patterns and limits on predictability. Ecol. Modell. 222, 2663–2675 (2011).
61
G. T. Evans, J. S. Parslow, A model of annual plankton cycles. Biol. Oceanogr. 3, 327–347 (1985).
62
B. Chen, H. Liu, Relationships between phytoplankton growth and cell size in surface oceans: Interactive effects of temperature, nutrients, and grazing. Limnol. Oceanogr. 55, 965–972 (2010).
63
W. S. Broecker, T.-H. Peng, Tracers in the Sea (Eldigio, Palisades, NY, 1982).
64
B. Mouriño-Carballido et al., Magnitude of nitrate turbulent diffusion in contrasting marine environments. Sci. Rep. 11, 18804 (2021).
65
K. Buck, F. Chavez, L. Campbell, Basin-wide distributions of living carbon components and the inverted trophic pyramid of the central gyre of the North Atlantic Ocean, summer 1993. Aquat. Microb. Ecol. 10, 283–298 (1996).
66
R. La Ferla et al., Vertical distribution of the prokaryotic cell size in the Mediterranean Sea. Helgol. Mar. Res. 66, 635–650 (2012).
67
R. La Ferla et al., Cell size and other phenotypic traits of prokaryotic cells in pelagic areas of the Ross Sea (Antarctica). Hydrobiologia 761, 181–194 (2015).
68
M. L. Calleja et al., The mesopelagic scattering layer: A hotspot for heterotrophic prokaryotes in the Red Sea twilight zone. Front. Mar. Sci. 5, 259 (2018).
69
A. Morel, Y.-H. Ahn, F. Partensky, D. Vaulot, H. Claustre, Prochlorococcus and Synechococcus: A comparative study of their optical properties in relation to their size and pigmentation. J. Mar. Res. 51, 617–649 (1993).
70
E. T. Buitenhuis et al., Picophytoplankton biomass distribution in the global ocean. Earth Syst. Sci. Data 4, 37–46 (2012).
71
B. Bergman, G. Sandh, S. Lin, J. Larsson, E. J. Carpenter, Trichodesmium: A widespread marine cyanobacterium with unusual nitrogen fixation properties. FEMS Microbiol. Rev. 37, 286–302 (2013).
72
H. N. Schulz, B. B. Jorgensen, Big bacteria. Annu. Rev. Microbiol. 55, 105–137 (2001).
73
D. A. Hutchins, D. G. Capone, The marine nitrogen cycle: New developments and global change. Nat. Rev. Microbiol. 20, 401–414 (2022).
74
D. G. Capone, J. P. Zehr, H. W. Paerl, B. Bergman, E. J. Carpenter, Trichodesmium, a globally significant marine cyanobacterium. Science 276, 1221–1229 (1997).
75
J. P. Zehr, D. G. Capone, Changing perspectives in marine nitrogen fixation. Science 368, eaay9514 (2020).
76
E. J. Carpenter, A. Subramaniam, D. G. Capone, Biomass and primary productivity of the cyanobacterium Trichodesmium spp. in the tropical N Atlantic ocean. Deep Sea Res. Part I Oceanogr. Res. Pap. 51, 173–203 (2004).
77
T. G. Boatman, G. J. G. Upton, T. Lawson, R. J. Geider, Projected expansion of Trichodesmium’s geographical distribution and increase in growth potential in response to climate change. Glob. Change Biol. 26, 6445–6456 (2020).
78
S. Bengtson, T. Sallstedt, V. Belivanova, M. Whitehouse, Three-dimensional preservation of cellular and subcellular structures suggests 1.6 billion-year-old crown-group red algae. PLoS Biol. 15, e2000735 (2017).
79
S. Zhang et al., Eukaryotic red and green algae populated the tropical ocean 1400 million years ago. Precambrian Res. 357, 106166 (2021).
80
E. J. Javaux, K. Lepot, The Paleoproterozoic fossil record: Implications for the evolution of the biosphere during Earth’s middle-age. Earth Sci. Rev. 176, 68–86 (2018).
81
E. N. Flannery, S. C. George, Assessing the syngeneity and indigeneity of hydrocarbons in the∼ 1.4 Ga Velkerri Formation, McArthur Basin, using slice experiments. Org. Geochem. 77, 115–125 (2014).
82
K. Nguyen et al., Absence of biomarker evidence for early eukaryotic life from the Mesoproterozoic Roper Group: Searching across a marine redox gradient in mid-Proterozoic habitability. Geobiology 17, 247–260 (2019).
83
G. M. Luo, C. Hallmann, S. C. Xie, X. Y. Ruan, R. E. Summons, Comparative microbial diversity and redox environments of black shale and stromatolite facies in the Mesoproterozoic Xiamaling Formation. Geochim. Cosmochim. Acta 151, 150–167 (2015).
84
M. Blumenberg, V. Thiel, W. Riegel, L. C. Kah, J. Reitner, Biomarkers of black shales formed by microbial mats, Late Mesoproterozoic (1.1 Ga) Taoudeni Basin, Mauritania. Precambrian Research. 196–197, 113–127 (2012).
85
Q. Luo, S. C. George, Y. Xu, N. Zhong, Organic geochemical characteristics of the Mesoproterozoic Hongshuizhuang Formation from northern China: Implications for thermal maturity and biological sources. Organic Geochemistry 99, 23–37 (2016).
86
E. Suslova, T. Parfenova, S. Saraev, K. Nagovitsyn, Organic geochemistry of rocks of the Mesoproterozoic Malgin Formation and their depositional environments (southeastern Siberian Platform) Russian Geology and Geophysics 58, 516–528 (2017).
87
K. L. French et al., Reappraisal of hydrocarbon biomarkers in Archean rocks. Proc Natl Acad Sci USA. 112, 5915–5920 (2015).
88
E. J. Javaux, A. H. Knoll, M. R. Walter, Morphological and ecological complexity in early eukaryotic ecosystems. Nature 412, 66–69 (2001).
89
L. Wu, A. Geng, Differences in the thermal evolution of hopanes and steranes in free and bound fractions. Org. Geochem. 101, 38–48 (2016).
90
Z. Li, H. Huang, S. Zhang, The effect of biodegradation on bound biomarkers released from intermediate-temperature gold-tube pyrolysis of severely biodegraded Athabasca bitumen. Fuel 263, 116669 (2020).
91
D. B. Mills et al., Eukaryogenesis and oxygen in Earth history. Nat. Ecol. Evol. 6, 520–532 (2022).
92
T. Fenchel, B. F. Finlay, Ecology and Evolution in Anoxic Worlds, R. M. May, P. H. Harvey, Eds. (Oxford University Press, Oxford, United Kingdom, 1995).
93
T. Fenchel et al., Microbial diversity and activity in a Danish fjord with anoxic deep water. Ophelia 43, 45–100 (1995).
94
T. F. Hansen, NUMmodel_Proterozoic_eukaryotes. github. https://github.com/trinefrisbaek/NUMmodel_Proterozoic_eukaryotes. Deposited 30 June 2022.
Information & Authors
Information
Published in
Classifications
Copyright
Copyright © 2022 the Author(s). Published by PNAS. This article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND).
Data, Materials, and Software Availability
Previously published data were used for this work (44). The model used can be found at: https://github.com/trinefrisbaek/NUMmodel_Proterozoic_eukaryotes (94).
Submission history
Received: December 26, 2021
Accepted: August 25, 2022
Published online: October 3, 2022
Published in issue: October 11, 2022
Keywords
Acknowledgments
We thank Karl Attard for his help with the graphical software and Shuhai Xiao for sharing his database on acritarch size distribution through the Proterozoic eon. We also acknowledge discussions with Susan Porter and Nick Butterfield. The work was supported by Villum Fonden Grant 16518 and by the Villum Kahn Rasmussen (VKR) Centre of Excellence “Ocean Life.”
Notes
Reviewers: S.X., Virginia Polytechnic Institute and State University; and E.Z., Carnegie Institution for Science.
Authors
Competing Interests
The authors declare no competing interest.
Metrics & Citations
Metrics
Altmetrics
Citations
Cite this article
A case for an active eukaryotic marine biosphere during the Proterozoic era, Proc. Natl. Acad. Sci. U.S.A.
119 (41) e2122042119,
https://doi.org/10.1073/pnas.2122042119
(2022).
Copied!
Copying failed.
Export the article citation data by selecting a format from the list below and clicking Export.
Cited by
Loading...
View Options
View options
PDF format
Download this article as a PDF file
DOWNLOAD PDFLogin options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginRecommend to a librarian
Recommend PNAS to a LibrarianPurchase options
Purchase this article to access the full text.