Space for hydrogen

Mats of photosynthetic microorganisms known as cyanobacteria generate molecular hydrogen. If they were doing so earlier in Earth's history, the effect on the evolution of the atmosphere could have been profound.

Hydrogen is an ideal energy source, not only for an industrialized society but also for various common bacteria. In the anoxic world of soils and sediments, molecular hydrogen is generated by the bacterial degradation of organic material. The hydrogen is immediately recycled and supplies much of the reducing power required for the biological formation of vast amounts of methane and hydrogen sulphide.

On page 324 of this issue1, Hoehler and colleagues show that mats of photosynthetic cyanobacteria growing on coastal mud flats are hot-spots of biological hydrogen production. Although the process is indirectly driven by solar energy, it takes place mainly at night when oxygen levels are low. The authors draw the astonishing conclusion that such a process might have contributed to the oxidation of the primordial Earth at a time when ancestors of modern microbial mats dominated the biosphere. Their reasoning is that, over time spans of a billion years or so, even a slow generation of hydrogen and its escape through the atmosphere and into space would result in a considerable overall loss of reducing power on the planet. The outcome would be that Earth's surface would eventually have become much more oxidized.

Mats of cyanobacteria (formerly called blue-green algae) consist of compact masses of microorganisms that grow on sediments and other solid surfaces in shallow aquatic environments (Fig. 1a)2. The penetration of daylight into the mat is often limited to the top millimetre, yet the gross photosynthesis achieved, as measured from the rapid evolution of oxygen, is impressive — up to 5 grams of carbon per square metre per day. This is comparable to the production in tropical rain forests. The net productivity of mats, however, is marginal because 99% of the biomass produced is rapidly remineralized by bacteria (Fig. 1b). The mats are excellent food sources for snails, crustaceans and other small invertebrates, so nowadays they form persistent communities only in extreme environments such as salt ponds or hot springs, where the extreme conditions exclude grazing organisms. An extensive record of fossil microbial mats (stromatolites) is evidence that in Precambrian time — before about 600 million years ago and the evolution of multicellular grazers — similar communities apparently covered most environments that had the appropriate levels of water and sunlight3.

Figure 1: Cyanobacteria, and their possible influence on Earth's early atmosphere.


a, Scanning electron micrograph of spaghetti-like, 3-μm-wide filaments of the microorganism Microcoleus chthonoplastes. Cyanobacteria such as these form dense mats that grow in shallow aquatic environments. Before the evolution of multicellular grazing organisms, such mats may have constituted the dominant biological community on Earth. b, The biogeochemical cycle in cyanobacterial mats. Almost all of the mats' photosynthetic biomass production is remineralized by bacteria back into carbon dioxide and water. The oxygen resulting from photosynthesis can accumulate only when reducing power in the form of organic carbon is buried or when molecular hydrogen is generated and escapes into space. The feature of especial interest here is the discovery by Hoehler and colleagues1 that cyanobacterial mats produce hydrogen as a by-product of nitrogenase activity during the night. Their hypothesis is that although hydrogen production is small relative to the photosynthesis–respiration cycle, over vast periods of geological time enough hydrogen has been lost to space to lower Earth's reducing capacity significantly and so raise oxygen levels in the atmosphere.

Today, hypersaline lagoons and intertidal flats such as those along the Pacific coast of Baja California, Mexico, provide ideal breeding grounds for cyanobacterial mats. Some, dominated by the species Microcoleus chthonoplastes, grow to many centimetres thick under the brine and have a finely laminated structure ranging in texture from jelly to rubber. Others, dominated by species of Lyngbya, persist on sand flats exposed to air, and form thin black carpets as tough as a doormat. Their coherent structure is created by intertwined microscopic filaments from billions of cyanobacteria embedded in gelatinous sheaths2. To gain protection from the burning sun and from harmful ultraviolet radiation, the cells secrete sunscreen pigments, which make the mat appear almost black.

Hoehler et al.1 studied hydrogen and carbon monoxide generation in both of these mat types, and discovered comparable physiological mechanisms of gas production in the two. The patterns of production were reversed, however, with high levels of carbon monoxide during the day, and high levels of hydrogen at night. The authors' larger argument hinges on hydrogen.

During the day, intense oxygen production in the millimetre-thin illuminated surface layer led to oxygen supersaturation and the development of permanent tiny gas bubbles trapped in the mats' fibres. The authors applied a simple but elegant technique to sample such bubbles repeatedly and analyse their gas composition throughout the day and night. Surprisingly, hydrogen concentrations in the Lyngbya mats were ten thousand times higher late at night than during the day. In Microcoleus mats, the hydrogen levels were also highest at night. But the day–night variation was much less pronounced, perhaps because, unlike Lyngbya, Microcoleus does not produce gas-reservoir bubbles.

Hoehler et al. propose that, in addition to its formation from fermentation processes in the anoxic mat, hydrogen is generated by the cyanobacteria themselves as a side reaction of their nitrogenase enzyme system. Nitrogenase occurs in many microorganisms, its main function being to 'fix' molecular nitrogen and convert it into nitrogen-rich cellular proteins. Marine ecosystems generally suffer from nitrogen starvation, so this process provides a crucial extra source of the nutrient.

Nitrogenase might, however, also be used to channel electrons into protons to form molecular hydrogen, a process known from planktonic cyanobacteria in lakes and in the sea4. The cyanobacteria are thereby relieved of excess reducing power accumulated during daylight in the form of storage carbohydrates in the cells. Nitrogenase is sensitive to oxygen, so most nitrogen-fixing cyanobacteria contain it in specialized cells, heterocysts, that lack the oxygen-generating photosynthetic machinery and have thick walls impermeable to external oxygen. Non-heterocystous cyanobacteria lack this oxygen-protection mechanism and therefore require a spatial or temporal separation of photosynthesis and nitrogen fixation. As a result, their nitrogenase is active mainly during the night5. The results of Hoehler et al. show that biogeochemists likewise had to be active during the night to discover this transient phenomenon.

It is well known that, in anoxic environments, a large fraction of the entire energy flow from organic matter is channelled through hydrogen, which is taken up by microorganisms and used to generate methane or hydrogen sulphide. Why, then, is the hydrogen production of cyanobacterial mats so exciting? There are two reasons. First, some of the hydrogen generated at the surface of the mats may evade rapid recycling by bacteria and escape into the atmosphere. Second, these mats have been the dominant biological communities for most of Earth's history, and their catalytic capabilities might therefore have had a profound influence on the chemical development of our planet.

Although oxygen-generating photosynthesis might have evolved more than 2.7 billion years ago, the concentration of oxygen in the atmosphere did not exceed 1–2% of the modern level until half a billion years later6,7. A net accumulation of oxygen on Earth requires that an equivalent quantity of reduced chemical species be taken out of the global cycles and stored away for long geological periods. The accelerating build-up of an oxygenated atmosphere and ocean over the Proterozoic era, from about 2.2 billion years onwards, was thus balanced by a burial of organic carbon and pyrite (iron disulphide) deep in marine sediments6. Loss of biologically produced hydrogen from the outer atmosphere into space could have been another mechanism for the removal of reducing power. Until now, however, this route has been considered to be of minor importance: only hydrothermal vents and volcanoes, or the ultraviolet-catalysed oxidation of atmospheric methane to hydrogen and carbon monoxide, have been considered as hydrogen sources.

Hoehler et al.1 propose that hydrogen generation by cyanobacterial nitrogenases might have been a further source, and one that had a global impact. Taking the results of their experiments on small pieces of mat from Baja California, they courageously extrapolate back to the first half of Earth's evolution. The gas budgets that they estimate are based on the assumptions that global primary productivity of microbial mats on the early Earth was comparable to that of modern marine ecosystems and that hydrogen emission rates were likewise similar to those studied. If the hydrogen were completely lost to space from an early atmosphere that was nearly devoid of oxygen, then this sink for a reducing species could have been as important as the deep burial of organic carbon for oxygenating Earth's surface.

There are still huge uncertainties and many open questions, of course. Were the primordial mats physiologically similar to those of modern Microcoleus or Lyngbya communities? What extent of the Earth's surface did they cover, and how productive were they? When did nitrogenase first occur in the ancestors of modern cyanobacteria? Nonetheless, in principle Hoehler et al. show that over a vast period of time the release of even a biological trace gas from a metabolic side reaction might have had profound consequences. With this work we can add a new piece of evidence to the great puzzle of early biogeochemical development on our planet.


  1. 1

    Hoehler, T. M., Bebout, B. M. & Des Marais, D. J. Nature 412, 324–327 (2001).

    ADS  CAS  PubMed  Google Scholar 

  2. 2

    Cohen, Y. & Rosenberg, E. (eds) Microbial Mats: Physiological Ecology of Benthic Microbial Communities (Am. Soc. Microbiol., Washington DC, 1989).

    Google Scholar 

  3. 3

    Schopf, J. W. & Klein, C. (eds) The Proterozoic Biosphere (Cambridge Univ. Press, 1992).

    Google Scholar 

  4. 4

    Conrad, R. Adv. Microb. Ecol. 10, 231–283 (1988).

    CAS  Google Scholar 

  5. 5

    Stal, L. J. New Phytol. 131, 1–32 (1995).

    CAS  Google Scholar 

  6. 6

    Des Marais, D. J., Strauss, H., Summons, R. E. & Hayes, J. M. Nature 359, 605–609 (1992).

    ADS  PubMed  Google Scholar 

  7. 7

    Nisbet, E. G. & Fowler, C. M. R. Proc. R. Soc. Lond. B 266, 2375–2382 (1999).

    Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Bo Barker Jørgensen.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Jørgensen, B. Space for hydrogen. Nature 412, 286–289 (2001).

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.