Energetics and population genetics at the root of eukaryotic cellular and genomic complexity

In PNAS, Lynch and Marinov provide detailed estimates of the energy cost associated with the addition of new coding or noncoding sequences to prokaryotic and eukaryotic genomes (1). The amount of ATP that is expended at each step of information transfer is derived from the known biochemistry of these processes and an extensive collection of data on gene expression, as well as nucleic acid and protein decay for diverse organisms. The energy cost is then transformed into fitness cost via a simple, intuitive notion that fitness cost is proportional to the fraction of the total energy expenditure of a cell that is attributable to the maintenance of a given sequence.

The laws of thermodynamics dictate that the stability and expansion of any physical system including, naturally, evolution of living organisms, are constrained by the adequate energy availability. Under the laws of population genetics that are deeply analogous to the laws of thermodynamics (2), the efficacy of selection in an evolving population is determined by the effective population size. In small populations, only mutations with a large phenotypic effect (selection coefficient) cross the barrier imposed by the genetic drift and are either eliminated from the population or fixed (depending on the sign of the selection coefficient), whereas in large populations, even mutations with a slight deleterious or beneficial effect are subject to efficient selection (3). Although this is rarely addressed in explicit terms, any evolutionary scenario can be considered seriously if and only if it falls within both the energetic and the population-genetic constraints (4).

Lynch and Marinov (1) show that the fitness cost of a new sequence negatively scales with the cell size, which is readily understandable because the energy cost does not necessarily strongly depend on the cell size, whereas the total energy expenditure is proportional to the cell volume. For a new sequence to be detected and eliminated by purifying selection from a haploid population, its cost should exceed 1/N_e (where N_e is effective population size). The estimates of Lynch and Marinov clearly show that this threshold is exceeded in prokaryotes, which accordingly cannot fix even a short sequence unless it provides a significant selective advantage, but not in eukaryotes that cannot eliminate a gene unless it is translated. Therefore, genome expansion in eukaryotes is possible under a neutral evolutionary regime and does not require any significant increase in energy production.

Energetic considerations appear to be particularly relevant when it comes to the origin of eukaryotes (eukaryogenesis), one of the major transitions in the evolution of cellular life that resulted in an unprecedented increase in the complexity of the cellular organization and genome architecture (5). All extant eukaryotic cells possess mitochondria, the power-generating membranous organelles that evolved from α-proteobacterial endosymbionts, or degenerated derivatives of the mitochondria, such as hydrogenosomes (5, 6). By inference, the last eukaryotic common ancestor already carried mitochondria. Lane and Martin have argued that the boost in energy production provided by the acquisition of (proto)mitochondria was a prerequisite for the evolution of the eukaryotic cell and, in particular, the complexification and drastic remodeling of the genome that accompanied eukaryogenesis (7, 8). In addition to the sheer expansion, this genomic transformation included insertion of multiple introns into the protein-coding genes, disruption of operons, and linearization of chromosomes, which became possible thanks to the recruitment of the telomerase from a prokaryotic mobile element (9). The estimates of Lynch and Marinov (1) show that in and by itself the eukaryotic genome expansion did not require the energy production boost from the mitochondria.

Does this demonstration imply that the acquisition of mitochondria was not the key factor, possibly a prerequisite, of eukaryogenesis? Not at all. The estimates indicate that eukaryotes can acquire DNA sequences of considerable length effectively for free, or put another way, are unable to eliminate invading DNA. This feature of eukaryotic biology apparently is a result of the synergistic effects of the large cell size that is characteristic of eukaryotes and results in the low fitness cost of new sequences (the energy expenditure for the maintenance of a new sequence is negligible compared with the total energy budget of a eukaryotic but not a prokaryotic cell) and the small population size that limits the power of selection (1). However, these estimates provide no clue as to how these intrinsic features of eukaryotic cells that seem to be prerequisites for the genome expansion could have emerged in the first place. Unlike genome expansion, increase of the cell size does require substantial amounts of extra energy per cell, which becomes problematic if only because the efficacy of membrane energetics [the only highly efficient form of biological energy conversion (10)] scales with the square of the cell diameter, whereas the cell volume obviously scales with the cube of the same. Membrane-bound energy-transforming organelles (i.e., mitochondria) offer the best known solution. In a broad agreement with the hypothesis of Lane and Martin (8, 10), the contribution of mitochondria to energy production could have enabled the drastic enlargement of the cells in the population that acquired the endosymbiont.

The large cell size in the primordial eukaryotic lineage leading to the last eukaryotic common ancestor is not a certainty. Indeed, numerous, phylogenetically diverse picoeukaryotes with cell sizes barely exceeding the typical prokaryotic size are known to exist

As shown by the estimates of Lynch and Marinov, a small population of large cells would have been conducive to the capture of substantial amounts of new DNA.

(11, 12). Nevertheless, it appears likely that picoeukaryotes independently emerged via reductive evolution in different lines of descent. Furthermore, early cell enlargement triggered by the mitochondria would be accompanied by a drop in N_e [given the apparently universal inverse correlation between the size of a cell or an organism and effective population size (3)], which would create the niche for genome expansion and remodeling.

The role of mitochondria in eukaryogenesis seems to extend far beyond the boost to energy production. Given the fundamental differences in the structure of the membrane phospholipids in archaea on the one hand, and bacteria and eukaryotes on the other hand (13), it appears likely that the mitochondria spawned the proliferation of the eukaryotic endomembranes, including the endoplasmic reticulum and the nuclear envelope along with the replacement of the plasma membrane with the bacterial version (14). Furthermore, because of the continuous lysis of the endosymbionts inside the evolving chimeric cell, the constant release of the endosymbiont DNA most likely triggered genome remodeling, via insertion of self-splicing group II introns into the host genes and transfer of numerous endosymbiont genes to the host genome (9, 14).

If the establishment of endosymbiosis that led to the emergence of eukaryotes was a singular occurrence, we may never know the actual chain of events. Nevertheless, analysis of energetic and population-genetic requirements for various evolutionary novelties, along with more specific biological evidence, constrain the possibilities and limit the space of plausible evolutionary scenarios. The above considerations suggest that, in the case of eukaryogenesis, evolution of cellular organization could have been the primary driver of genomic innovations, not the other way around (Fig. 1). The recent discovery of Lokiarchaeota (Loki), the archaeal sister group of eukaryotes that are indistinguishable from other archaea in terms of genome size and architecture but encode a variety of proteins indicative of complex intracellular organization, in particular an advanced cytoskeleton (15–18), appears best compatible with the possibility that the host of the proto-mitochondrial endosymbiont was an archaeon, albeit one with an unusually elaborate intracellular organization (Fig. 1). Engulfment and domestication of an α-proteobacterium by a Loki-like archaeon could have led to substantially increased energy production per (chimeric) cell, which would trigger cell enlargement accompanied by intracellular compartmentalization. As shown by the estimates of Lynch and Marinov, a small population of large cells would have been conducive to the capture of substantial amounts of new DNA (1). Together with the onslaught of the endosymbiont DNA, this condition could have led to genome remodeling and expansion, molding the eukaryotic genomes (Fig. 1).

Fig. 1.

Cellular organization-driven evolutionary scenario of eukaryogenesis. Eukaryogenesis is represented by three stages: 1, α-proteobacterium engulfed by a Loki-like archaeon becomes the proto-mitochondrial endosymbiont; excess energy production; 2, excess energy production leads to cell enlargement accompanied by intracellular compartmentalization (emergence of endomembranes) and a drop in N_e; 3, expansion and remodeling of the genome including insertion of endosymbiont DNA (both genes and introns) accompanied by shrinking of the endosymbiont genome and linearization of the chromosomes; further evolution of intracellular organization including the origin of the nucleus from the endomembranes, conceivably, as a protective device against translation of aberrant, intron-containing transcripts (14). The cellular complexity of the Loki-like putative host of the endosymbiont is unknown; the figure only includes some inferences from the gene repertoire of the Lokiarchaeon. Abbreviations: CC, circular chromosome; CS, cytoskeleton; EM, endomembranes; LC, linear chromosomes; N, nucleus.