Evolutionary biology

Oxygen at life's boundaries

Article metrics

Proteins are made of amino acids. But amino acids are made of atoms. Exploration of this self-evident principle opens up fresh perspectives on the evolution of biological membranes and multicellular life.

For many microorganisms, one cell is adequate; for some plants and animals, billions are scarcely enough. But whatever the number, the cell is the fundamental unit of living matter, and is invariably delineated by a membrane — the plasma membrane — that is a selective barrier separating the inside from the outside. Some cells may also contain compartments, which are bounded by further membranes. Communication between intracellular compartments, or between cells and their environment, relies on transmembrane proteins that span the entire biological membrane. Using the unfamiliar prism of atomic rather than amino-acid composition, Acquisti et al.1 show how their inspection of all the transmembrane proteins of 19 contemporary organisms tells us a lot about evolution. Their results appear on page 47 of this issueFootnote 1.

Cells are divided into two large groups: eukaryotic, in which the DNA molecules are bounded by a nuclear membrane; and prokaryotic, which have no nuclear membrane. Prokaryotes are never found as complex, multicellular organisms. And whereas prokaryotes possess only simple intracellular compartments, or none at all, all eukaryotic cells contain compartments that are surrounded by two membranes. So understanding how and when compartmentalized cells appeared on Earth is one of the big questions in biology, as is understanding how and when multicellular eukaryotic organisms emerged millions of years later. Acquisti et al.1 provide novel evidence of the absolute requirement of atmospheric oxygen (O2) for these transitions to happen.

The 'oxygen revolution' stems from the first appearance, 3 billion years ago, of organisms releasing O2 as a metabolic waste. This process led to a first great 'oxygenation event', 800 million years later, with a second one occurring one billion years ago. This second event is believed to have eventually fuelled the appearance of complex life-forms during the Cambrian explosion about 543 million years ago2. More recently, 425 million years ago, O2 levels were a major factor in the progressive adaptation of aquatic arthropods and vertebrates to terrestrial life3. Accordingly, evolutionary analyses encompassing the past 2.3 billion years have revealed a correlation between increased organism complexity and the development of aerobic metabolism4.

Two explanations have been given for this correlation, both invoking metabolic fitness. The first is that, compared with their anaerobic ancestors, oxygen-respiring cells are highly efficient energy-extracting machines: cells can use O2 as an electron acceptor in respiration processes, and because of its high reduction potential, the maximum energy can then be released from nutritional resources. A second, complementary explanation stems from the observation that O2 allows a thousand more metabolic reactions than can occur in anoxic conditions5,6.

Acquisti et al.1 now propose a third explanation, this time based on functional constraints. They argue that, in low O2 conditions, it was impossible for cells to synthesize or maintain novel communication-related transmembrane proteins. Such proteins would be required for intracellular compartments to work together, a prerequisite to compartmentalization. Because evolution from unicellular to multicellular organisms requires efficient communication between cells, this evolutionary step was similarly hindered by insufficient levels of O2. Surprisingly, Acquisti and colleagues' analyses suggest that the main distinctive feature of these novel transmembrane proteins is that they are enriched in oxygen atoms: in particular, their oxygen-rich external domains are longer than those of transmembrane proteins from uncompartmentalized cells.

How might levels of atmospheric oxygen have constrained the atomic composition of transmembrane proteins? Acquisti et al. propose two hypotheses. The first, inspired by stoichiometric ecology7, is that, in the absence of O2, building oxygen-rich amino acids would have been too demanding. However, there is no obvious evidence for such a metabolic limitation. According to computational analyses5, the seven amino acids containing most oxygen atoms — D, E, Y, S, T, N and Q, in single-letter code — could all be synthesized by anoxic metabolisms. Moreover, among the 86 final reactions producing these amino acids, only 11 are specific to the oxic metabolism5,8. The authors thus favour the second hypothesis: that the reducing atmosphere found under low levels of O2 would have damaged long, oxygen-rich protein domains, and made the synthesis of transmembrane proteins with long external parts impractical.

The work of Acquisti et al.1 could be refined by correlating the oxygen content of transmembrane proteins with that of the compartments in which they are embedded. In particular, intracellular membranes delineating reducing compartments would be expected to contain more oxygen-poor proteins than does the plasma membrane. A candidate for such a study would be the membrane of mitochondria, the cell's energy-producing compartments. As a result of their continuous consumption of oxygen by respiration, mitochondria are the most anoxic compartments of oxygen-respiring cells.

Moreover, a striking difference between most eukaryotes and most prokaryotes is that respiration does not occur in the eukaryotic plasma membrane. As a consequence, O2 is not consumed in the immediate proximity of eukaryotic plasma membranes, which thus exist in an oxygen-rich environment. Seen in the light of Acquisti and colleagues' paper, this may have been a factor in protecting their oxygen-rich transmembrane proteins. The O2-driven emergence of multicellular organisms may therefore have required two major changes: accumulation of oxygen-rich proteins in the plasma membrane and confinement of respiration to intracellular compartments dedicated to that purpose.

It is striking that it has taken so long to make these simple observations on the elemental compositions of transmembrane proteins, and to formulate the resulting model of evolution1. The main reason may be that biologists usually regard proteins as chains of amino acids, or combinations of polypeptide domains, and ignore the fact that, in essence, proteins are arrangements of atoms. The elemental structure of biopolymers may well have been shaped by nutritional, physical or functional constraints9, but the effects of these constraints usually remain hidden if one inspects only the amino-acid (or base-pair) compositions.

The work of Acquisti et al. is a welcome reminder that such constraints acted on subsets of proteins linked by function10, cellular location1 or metabolic role11, as well as on the total protein content of a cell12. Structural biologists are no longer alone in keeping the atomic composition of their favourite proteins under close scrutiny — evolutionary biologists, too, will find this a fruitful pursuit.


  1. 1.

    This article and the paper concerned1 were published online on 20 December 2006.


  1. 1

    Acquisti, C., Kleffe, J. & Collins, S. Nature 445, 47–52 (2007).

  2. 2

    Knoll, A. H. Life on a Young Planet: The First Three Billion Years of Evolution on Earth (Princeton Univ. Press, 2003).

  3. 3

    Ward, P., Labandeira, C., Laurin, M. & Berner, R. A. Proc. Natl Acad. Sci. USA 103, 16818–16822 (2006).

  4. 4

    Hedges, S. B., Blair, J. E., Venturi, M. L. & Shoe, J. L. BMC Evol. Biol. 4, 2 (2004).

  5. 5

    Raymond, J. & Segrè, D . Science 311, 1764–1767 (2006).

  6. 6

    Falkowski, P. G. Science 311, 1724–1725 (2006).

  7. 7

    Sterner, R. W. & Elser, J. J. Ecological Stoichiometry: The Biology of Elements from Molecules to the Biosphere (Princeton Univ. Press, 2002).

  8. 8


  9. 9

    Baudouin-Cornu, P. & Bragg, J. G. in Encyclopedia of Genetics, Genomics, Proteomics and Bioinformatics (eds Jorde, L., Little, P., Dunn, M. & Subramanian, S.) Section 3.3 (doi:10.1002/047001153X.g303318) (Wiley, New York, 2006).

  10. 10

    Mazel, D. & Marlière, P. Nature 341, 245–248 (1989).

  11. 11

    Baudouin-Cornu, P., Surdin-Kerjan, Y., Marlière, P. & Thomas, D. Science 293, 297–300 (2001).

  12. 12

    Elser, J. J., Fagan, W. F., Subramanian, S. & Kumar, S. Mol. Biol. Evol. 23, 1946–1951 (2006).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Baudouin-Cornu, P., Thomas, D. Oxygen at life's boundaries. Nature 445, 35–36 (2007) doi:10.1038/nature05521

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.