Metabolic evolution and the self-organization of ecosystems

Abstract

Metabolism mediates the flow of matter and energy through the biosphere. We examined how metabolic evolution shapes ecosystems by reconstructing it in the globally abundant oceanic phytoplankter Prochlorococcus To understand what drove observed evolutionary patterns, we interpreted them in the context of its population dynamics, growth rate, and light adaptation, and the size and macromolecular and elemental composition of cells. This multilevel view suggests that, over the course of evolution, there was a steady increase in Prochlorococcus' metabolic rate and excretion of organic carbon. We derived a mathematical framework that suggests these adaptations lower the minimal subsistence nutrient concentration of cells, which results in a drawdown of nutrients in oceanic surface waters. This, in turn, increases total ecosystem biomass and promotes the coevolution of all cells in the ecosystem. Additional reconstructions suggest that Prochlorococcus and the dominant cooccurring heterotrophic bacterium SAR11 form a coevolved mutualism that maximizes their collective metabolic rate by recycling organic carbon through complementary excretion and uptake pathways. Moreover, the metabolic codependencies of Prochlorococcus and SAR11 are highly similar to those of chloroplasts and mitochondria within plant cells. These observations lead us to propose a general theory relating metabolic evolution to the self-amplification and self-organization of the biosphere. We discuss the implications of this framework for the evolution of Earth's biogeochemical cycles and the rise of atmospheric oxygen.

Keywords: Earth history; Prochlorococcus; metabolic evolution; microbial oceanography; mutualism.

Fig. 1.

Typical relative abundance distributions of Prochlorococcus ecotypes as a function of depth and accompanying light intensity and nutrient concentration profiles in stratified oceanic waters. Ecotype populations are geographically and temporally dynamic, but in warm, stable water columns return to this same depth-differentiated state (12). The deepest branching ecotypes are most abundant at the bottom of the euphotic zone, where nutrient concentrations are high and light energy low. The most recently diverging ecotypes are most abundant near the surface, where the reverse is true (–14). HL, high-light-adapted; LL, low-light-adapted.

Fig. 2.

Illustration of approach to metabolic reconstructions. Phylometabolic trees reflect the evolution of metabolic network phenotypes because they integrate constraints from phylogenetics and metabolism (15, 16). All sequenced genomes within a given clade are searched for the presence/absence of enzymes catalyzing the reactions of different pathways. Mapping pathway variability patterns onto phylogenies of the clade suggests the order of metabolic innovations. In this example, three alternative pathways (pink, yellow, and blue) connect essential and universal pathways (black). Genes for the yellow pathway are nearly universally distributed (Inset), suggesting that it is the ancestral pathway, with the pink and blue pathways deriving from it. Maintaining continuity of flux results in trees of functional phenotypes. Biochemical differences between alternative pathways (e.g., ATP/trace metal requirements or oxygen sensitivities of their enzymes) suggest evolutionary driving forces (15, 16).

Fig. 3.

Metabolic evolution of Prochlorococcus. The metabolic variants are represented in simplified form along a gradient of the ratio of electron flux to nutrient flux ${\nu }_e/{\nu }_n$. Innovations (SI Appendix, Fig. S1 and Table S1) are highlighted with dashed orange lines. Cellular electron drains are represented as red dots and uptake pathways as blue dots. Photosystems are represented as boxed 2’s and 1’s, and their colors reflect the major absorption wavelengths, their line thickness reflects their absorption cross-section, and their relative size reflects the PSII:PSI ratio (SI Appendix, Fig. S1). Key innovations in the HLI and HLII clades are related to protection/repair of direct photodamage (52, 53), including UV damage, as indicated by a darkening protective pink shade.

Fig. 4.

The emergence of Prochlorococcus ecotypes. Eq. 5 suggests that innovations increasing ${\nu }_e/{\nu }_n$ trigger the emergence of new ecotypes (purple) that draw down limiting nutrients ($n$, green depth profile) in oceanic surface waters as they go through adaptive radiation. Nutrient drawdown near the surface increases ecosystem biomass and restricts ancestral ecotypes (pink) with a higher minimum subsistence nutrient concentration $[n]$ to greater depths in the water column where nutrient levels remain higher. Finally, Eq. 5 suggests that by excreting increasing amounts of fixed carbon, while maximizing ${\nu }_e/{\nu }_n$, the evolution of Prochlorococcus has increased the long-term steady-state concentrations of DOC (orange depth profile).

Fig. 5.

Metabolic evolution of SAR11. Variants are shown in simplified form along a gradient of ${\nu }_e/{\nu }_n$. Dashed orange lines highlight innovations (SI Appendix, Fig. S4 and Table S2), which include disruptions or replacements of glycolysis (all branches), the step-wise completion of the glyoxylate shunt (from group V to IIB to all groups IA) and the gain and loss of excretion/uptake pathways. See main text for details. Excretion pathways are represented as red dots and uptake pathways as blue dots. This reconstruction is less certain than for Prochlorococcus because SAR11 has greater diversity (99, 100), and as highlighted in the phylogeny (included clades in green), many clades have no or few sequenced representatives (SI Appendix, Table S2). Several branches were therefore drawn as a polytomy.

Fig. 6.

The metabolic organization of Prochlorococcus and SAR11 and that of green plant cells. For simplicity, many reaction sequences are shown as single arrows without accurate stoichiometry. Blue/orange arrows indicate organic carbon flux through similar pathways between Prochlorococcus and SAR11 and between chloroplasts and mitochondria. The loss of glycolate uptake in the open ocean 1A.3 clade (Fig. 5) is indicated by a dashed arrow. The loss of catalase in Prochlorococcus is indicated by a crossed-out KatG gene, while this cross-out is dashed for SAR11 to indicate the loss of catalase in some of its later branching clades. In plant cells, catalase is similarly located in the peroxisome and not chloroplasts/mitochondria (–113) as indicated by the crossed-out KatG genes. PTOX and AOX provide electron drains in the electron transport chains of both systems, and chloroplasts and Prochlorococcus both use chlorophyll b as well as a (6). 2OG, 2-oxoglutarate; 3PG, 3-phosphoglycerate; ACE, acetyl-CoA; KatG, catalase; CIT, citrate; DHAP, dihydroxyacetone phosphate; GLA, glycerate; GLC, glycolate; GLY, glycine; GOX, glyoxylate, ICE, isocitrate; MAL, malate; OAC, oxaloacetate; PEP, phosphoenoylpyruvate; PYR, pyruvate; RBP, ribulose bisphosphate; SER, serine; SUC, succinate.