Les plus anciennes formes de vie sont datées de près de 4 milliards d'années

Les roches les plus anciennes trouvées jusqu'aujourd'hui sur la terre sont les gneiss d'Acasta au nord-ouest du Canada près du Great Slave Lake (4.03 Ga), l'Isua Supracrustal au Groenland occidental (3.7 à 3.8 Ga). Mais des roches bien connues et presque aussi anciennes ont également été trouvées dans la vallée du Minnesota et au nord du Michigan (3.5-3.7 Ga), ainsi qu'au Swaziland (greenstones de Barbeton 3.4-3.5 Ga ), et en Australie occidentale (3.4-3.6 Ga Apex Chert à Marble Bar, Western Australia).

Ces roches très anciennes, sont célèbres dans le monde scientifique parcqu'elles contiendraient des bactéries fossiles, c'est à dire des restes de forme de vie à un stade précoce de l'évolution de la vie sur la terre. Bien qu'il y ait contestation par certains que ces prétendus restes soient effectivement des fossiles d'organismes vivants primitifs (en raison du degré élevé de métamorphisme ayant affecté ces formations, une majorité semble néanmoins se dégager en faveur des fossiles. Quelque soit la réponse, il semble que le consensus scientifique actuel soit pour une origine très lointaine de la vie, peu de temps après la formation du système solaire, des planètes et de la terre il y a 4.5 Ga.

L'histoire des bactéries fossiles et les débats scientifiques les concernant sont donnés ci-après. Mais voir d'abord le site de l'Université de Münster en Allemagne, qui fait une présentation des bactéries fossiles et des sites où ils ont été découverts.

Comme certains liens de ce site sont inopérants,voir aussi le site de University of California

Fossilised bacteria in oldest rocks of the earth

Back in the 1980's a young scientist from the University of California at Los Angeles came to this blighted place in search of fossils. Bill Schopf was interested in the history of Precambrian life, that long era of Earth history spanning the Achaean and Proterozoic - each of which dwarfs the duration of the Phanerozoic that started some 550 million years ago with the rise of conventional fossils.

Schopf was laughed at by his colleagues, for how could you expect to find fossils in rocks that were so old they surely predated the evolution of bacteria? But Schopf was a determined man: he'd been involved in describing the unassailable bacterial fossils of the Gunflint Chert in northern North America and these were a staggering 2 billion years old (2Ga). These organisms had lived near the beginning of the Proterozoic, the younger of the Precambrian's two aeons (see fig 1). And if bacterial fossils of 2 Ga could be found, why not older?

The importance of this question could not be overstated because it brings us nearer to knowing when life on Earth began. Schopf knew well that there were only two places in the world where rocks of this age that might contain fossils were likely to be found; in the rocks of the Australian Pilbara Supergroup or in the rocks of the hills and valleys of the Barbeton Mountain Land of South Africa, the Swaziland Supergroup. Both of these sedimentary terrains have been pressure-cooked to some extent by post-depositional heating and sinking within the Earth's crust but the Barbeton Range has been cooked longer and hotter than the rocks of the Pilbara. This decided Schopf, he would make the best of this bad hand and travel to Australia to look for his fossils.

The outcome of his studies was the publication in Science in 1993 of a description of eleven microfossils that were claimed to be the oldest fossils in the world. They were all thought to be members of the cyanobacteria (the blue-green algae). The news was sensational at the time was because age-dating of the encapsulating rock units (using the very accurate uranium-lead technique, see Searching for Realtime in the January issue of Chemistry in Britain) constrained the age of the fossils to about 3.5 Ga - to put this into perspective remember that the dinosaurs became extinct 65 million years ago! The clear implication of this date was that life must have got started very soon indeed after the Earth formed (at 4.5 Ga) to have produced something as complex as single celled organisms that left morphological traces of themselves in the fossil record.

Because rocks of this Achaean age have all been metamorphosed to some extent, and since cyanobacteria don't have much morphology to start with, experts often disagree whether traces in the rock are truly fossil or simply a mineral artefact. The recent controversy over the 'nanobacteria' found in the Martian meteorite ALH84001 makes this point all too clearly. In the case of the search for life at this extremity of time these morphological clues are already right at the edge of credibility and will not do if we wish to extend the search to older rocks. But chemistry provides an answer that is considered definitive: the presence in the rock of isotopically depleted carbon. This technique relies on the fact that the Rubisco enzyme system of photosynthesis preferentially incorporates the light isotope of carbon (carbon-12) over the heavy isotope (carbon-13). This in turn is a consequence of the fact that carbon-12 is more kinetically active than carbon-13 and so encounters the binding site on the Rubisco enzyme more frequently.

Since photosynthesis is a very ancient biochemical trick that is common to simple unicellular as well as multicellular organisms it is reasonable to assume that if isotopically light carbon is found in ancient rocks then this is evidence for photosynthesis. At first sight is seems disarmingly simple to use something a trivial as a negative carbon isotope ratio to infer something as massive as the presence of life on Earth. But the truth is that there still seems to be no unequivocal way of inorganically mimicking the negative carbon isotope compositions that go hand-in-hand with the presence of life.

There is one place - and one place alone - where there are sedimentary rocks that are older than those of the Pilbara or Barbeton rocks and these are the Isua supercrustal rocks of western Greenland. They are a staggering 3.8 billion years old and naturally have been the target for investigating life's earliest history: in fact their investigation predates that of the Apex Chert fossils. But the Isua supercrustals have one big problem: they have been seriously pressure-cooked. Estimates suggest that within a billion years of deposition these sediments were subjected to temperatures in excess of 500 degrees C and pressures of greater than 5,000 atmospheres!

In short any traces of a bacterium or blue-green algal cell's morphology will have been totally wiped out. All we can do to check for the presence of life is to use the carbon isotope assay. This is truly testing for life at the edge of time for the Isua rocks are almost 400 million years older than the fossils of the Apex Chert; that is about the same amount of time that separates humanity from the doughty lung-fish that decided to make the leap out of the ocean and try their hand at living on land.

The carbon test was applied to the Isua rocks in the late 1970's. It was a bulk isotopic analysis of the carbon in the rock. 'Bulk' simply means that the entire sample was homogenised and analysed, necessary in the days before mass spectrometers became refined enough to work on truly tiny samples. The results were equivocal: only a little more negative than the values that could be caused by inorganic carbon isotope fractionation (see Fig 2), but for true believers they were hailed as evidence that life had existed as early as 3.8 Ga.

But technology moves on and by the mid-1990's it was no longer the case that an entire sample had to be homogenised to get enough carbon to measure isotopically. It was now possible to use an ion-microprobe to direct a beam at specific particles within the rock, ionise discrete carbon inclusions and measure their isotopic composition by magnetic-sector mass spectroscopy. This was the technique used by a team led by Stephen Mojzsis in California in the mid-1990s.

Mojzsis' team focused not merely on the Isua rocks but a specific enclave that crops out on the offshore island of Akilia, where the rocks are a staggering 3.85Ga - fifty million years older than the surrounding rocks. The team made multiple analyses of picogram sized particles of carbon within the rock and discovered that without exception they were all isotopically depleted i.e. enriched in carbon-12. Moreover the ratios were far from the range diagnostic of inorganically derived carbon (fig 2) strengthening the case that they could only be the spoor of ancient photosynthesis. In addition the carbon globules were encased within apatite grains - the mineral building block of bones and teeth. Although apatite can be formed inorganically its presence in the rock with isotopically depleted carbon was doubly suggestive of the presence of life.

The dating of these Akilia Island samples was more controversial than that of the rest of the Isua complex. Even so there was one extraordinary and inescapable finding; it seemed certain that life evolved during, rather than after, the period of Earth history known as the 'late heavy bombardment' period of the Hadean Era, the era immediately prior to the Archaean. The Hadean ('hell-like') is so-called because at this time the Earth and its sister planets - the Moon and Mars - were still being bombarded by debris left over from the formation of the solar system. Curiously enough we know this by dating rocks from lunar craters returned by the Apollo missions - all of the Earth's own craters of Hadean age have long since been subducted and recycled by the motions of its tectonic plates. The late heavy bombardment finished as far as we can tell about 3.8Ga - the same age as the Isua rocks which supposedly already contained good evidence of life. It followed therefore that photosynthesis must have evolved a good while before this; during the late heavy bombardment period.

Life evolving under these circumstances is quite different to the way that Darwin envisaged it happening, in a 'warm little pool' of favourable prebiotic compounds and a few aeons of time for some gentle chemical cookery.

Indeed it seems probable that if these Akilia carbon inclusions are the spoor of life then it evolved somewhere away from the surface bombardment, perhaps in the depths of the ocean near hydrothermal vent systems.

New genetic evidence from contemporary bacteria that inhabit these environments seems to indicate that their biochemistry is very ancient indeed; could these perhaps be our own ultimate ancestors? It has also been suggested that wherever life evolved on this hellish Earth it may have needed several tries to get going, being repeatedly snuffed out by marauding asteroids that made the impact that did for the dinosaurs look as delicate as a falling thistledown.

But there is a new controversy over the rocks of Akilia Island. Central to the Mojzsis team's findings is the 'geological context' from which they took their samples, the certainty that no matter how cooked they are today these rocks were once sedimentary (part of a Banded Iron Formation, BIF). But even this most basic assumption has been challenged. In the 24 May 2002 issue of Nature Christopher Fedo and Martin Whitehouse argue that the rocks that the Mojzsis team took their samples from are not sedimentary BIFs at all but rather are igneous rocks - formed by the action of ancient volcanism. If this is the case then all bets are off and we need to re-examine the certainty that only organic chemical reactions can result in very negative carbon isotope ratios. We are also back to Schopf's bugs as the oldest evidence for life on Earth.

Or are we? The detailed Isua and Akilia isotopic analyses were all inspired by Schopf's finding of the Apex Chert fossils so it is doubly ironic that it is these themselves that are now the subject of renewed controversy. The evidence unseating the Apex Chert's claim to contain our planet's oldest fossils is once again the 'geological context' from which the samples came and the interpretation of carbon-isotope chemistry, albeit investigated by a different analytical technique, laser-Raman spectroscopy.

When a laser beam of a particular wavelength hits a target material, most of the light that bounces off the material remains the same wavelength. However a small proportion of the scattered laser light changes wavelength and the magnitude of the wavelength shift is determined by the molecular composition of the targeted material. This is the Raman effect. It turns out that every type of molecule has a unique signature. So laser-Raman spectroscopy is a step beyond the ion-microprobe. It can yield the molecular composition of materials, including fossils.

Martin Brasier and colleagues at Oxford University recently re-examined the Apex Chert samples and presented evidence suggesting that the structures within are all mineral artefacts. The structures are branched, Brasier and his colleagues insist - a morphology that is inconsistent with Schopf's claim that they are bacteria. Furthermore the Brasier team presents geological evidence showing that the rocks of the Apex Chert are criss-crossed by lava flows of the same age as the supposed sediments, negating Schopf's claim that the rocks were deposited under quiet near-shore conditions.

In the opposite camp, and in the same issue of Nature, the Schopf team presented isotopic evidence (using the laser-Raman technique) that seems to show conclusively that the fossils are all isotopically depleted. But Brasier and colleagues - who also employed Raman spectroscopy - conclude that the carbon is isotopically depleted not as a result of photosynthesis but instead because of a set of chemical reactions known as Fischer-Tropsch synthesis. Here carbon monoxide and hydrogen are catalysed to more complex carbon compounds at high temperatures in the presence of catalysts such as iron oxide.

This reaction may be accompanied by a shift to negative carbon isotope ratios (there is no clear consensus yet) that could mimic the isotopically negative signature supposedly uniquely indicative of life. Indeed there is evidence to suggest that the Fischer-Tropsch type synthesis is responsible for the negative carbon isotope ratios found in certain carbonaceous meterorites, an interesting parallel given that the search for the most ancient life on Earth requires the same types of techniques needed to identify the presence of life on other planets.

At the present time it is hard to disentangle the competing claims and counter claims for true photosynthetic fractionation versus the Fischer-Tropsch reaction in the Apex Chert debate. At first sight it seems strange to invoke a particular type of inorganic chemical reaction when three decades of previous interpretation have all favoured the isotopically depleted carbon -photosynthesis link.

However it is clear that the difficulties involved in unambiguously identifying life at these edges of time and space are extreme. The quality of evidence must be unassailable - this is not now the case with the Isua, Akilia and Apex Chert fossils. Time of course will sort out the Brasier-Schopf and Mojzsis-Fedo controversies and to the victor the spoils. But a more important question than academic bickering remains the origins of life itself. Did it really get snuffed several times before finally catching on? And was there one discrete moment, perhaps captured in the rocks of Akilia Island, where, to paraphrase my favourite Australian road-movie, 'it learned to live again?'

The Schopf Brasier debate on fossilised bacteria

A dispute over the identification of extremely ancient fossils is threatening to revise how scientists write the history of early life on Earth. In the journal Science, on April 30, 1993, J. William Schopf, a professor of paleobiology at the University of California, Los Angeles, described 3.5 billion-year-old specimens from Western Australia as fossilized bacteria.

The tiny organisms were encased in chert, a fine-grained sedimentary rock. On the basis of morphology (shape), Schopf sorted the bacteria into 11 taxa, or distinct groupings, and said 7 were probably early relatives of cyanobacteria, primitive photosynthetic organisms.

Cyanobacteria are credited with the first large-scale production of atmospheric oxygen. By converting a reducing (oxygen-poor) atmosphere to an oxidizing (oxygen-rich) one, cyanobacteria permitted the evolution of more advanced, oxygen-breathing organisms.

Yet the first geochemical evidence of significant atmospheric oxygen – evidence preserved in the chemical makeup of ancient rocks – came from about 1 billion years later.

Schopf’s identification, which gained acceptance as the first evidence of oxygen-producing organisms, is now under bruising attack. Critics charge not only that the structures Schopf observed in the rocks are not fossilized cyanobacteria, but also that they are not biological fossils at all.

Martin Brasier, professor of paleobiology at Oxford University, along with an international team, raised numerous counter-arguments in the March 7, 2002, edition of Nature. Brasier says Schopf misunderstood the geology of the supposed microfossils, which were preserved not in marine sediments, but rather in a hydrothermal vent – or even in volcanic glass.

"The context for the Apex chert 'microfossils' can be reconstrued as the subsurface feeder dyke to a submarine (possibly quite deep) hydrothermal vent, closely associated with volcanicity." (In other words, they formed in association with hot fluids near a volcanic structure.) The hot, low-oxygen conditions may have been "excellent for the prebiotic synthesis" of organic compounds, Brasier adds, but at several tens of meters below the vent, were "not a happy habitat for oxygen-producing cyanobacteria."

In an email, Brasier also charged that Schopf overstated or misinterpreted the regularity in the shape of his specimens. "Eleven fossil taxa (the greatest diversity in the fossil record until the Gunflint chert, some 1600 million years younger) were … based upon structures that are much more chaotic and inconsistent than shown and described."

This business of shape is important because Schopf used physical appearance as the basis of his original identification. Brasier says his examination of the specimens revealed structures "indistinguishable from graphitic mineral growths that occur alongside [the purported fossils] - they frequently branch chaotically in the same way and intergrade continuously with the shapes of inanimate matter."

Appearance is a poor guide to biogenic origin, Brasier concludes. "This provides a clear warning to all astrobiologists that complex, microfossil-like structures can easily arise from simple, abiogenic causes."

Only by carefully comparing non-biological and biological structures, he says, will it be possible to distinguish structures made by biology.

Schopf is having none of this. For starters, he calls the argument over cyanobacteria a "red herring. I have referred to these specimens as 'cyanobacterium-like,' 'possible cyanobacteria,' and even 'probable cyanobacteria,' but I have never, ever claimed that they were cyanobacteria per se. Nobody goes to look at the literature. I defined these taxa as being of uncertain systematic position. I said they were prokaryotes [microorganisms without cell nuclei], and still think they are, but I didn't know for sure then, and I don't know now, precisely what kinds of bacteria they are."

As for Brasier’s contention that fossils could not be found in a hydrothermal or volcanic structure, Schopf says that, while the conditions of deposition remain contentious, "It’s perfectly fine for there to be organisms in hot springs. Hot springs today are teeming with life; think of Yellowstone. If there was life around [3.5 billion years ago], it would be astounding that it was not present in such a hot-spring environment."

Schopf says his scientific approach is similar to that of Georges Cuvier, a founder of paleontology who first applied the Linnaean classification system to fossils. "Cuvier said that the way you solve the problem [of determining whether or not a specimen is biogenic] is that you identify a suite of traits that are unique to life, as a suite, and that not exhibited by inanimate matter, but are exhibited by living organism and the fossils."

The Apex chert specimens satisfy critical criteria for fossils, Schopf continues. "Every test that can be applied has been applied, and the Apex fossils meet them all, a suite of traits that taken together are simply unknown in nonbiologic materials."

Here are his arguments:

  1. Morphology: 200 specimens from the Apex rocks and numerous specimens from other rocks of similar age show the characteristic shapes of bacteria.
  2. Consistency: The fossils make sense in the context of other information about the epoch, and are not, for example, unduly advanced or retarded compared to other evidence from the same period. "There are hundreds of Precambrian fossil-bearing chert units like the Apex, and they have the same sorts of fossils," he says.
  3. Isotopic analysis: 10 analyses of the Apex charts, and 125 analyses on other fossil units of about the same age "all show the same distinctive biological [carbon isotope] signal."
  4. Composition: Raman imagery of 15 specimens from the Apex and other cherts of the same age, Schopf says, show the remains of organic matter. (Raman imagery is an analytical technique that can map the distribution of a particular chemical throughout a sample. It enables one to compare the location of specific chemical signals to visible shapes within the sample.

But not everyone is convinced by Schopf’s data. At a recent Geological Society of America conference, Jill Pasteris, a professor of earth and planetary sciences at Washington University who has used Raman spectroscopy for 20 years, presented a poster showing that Schopf’s conclusion about kerogen was "erroneous."

Kerogen is a grab-bag description of the remains of organic matter as it gradually decomposes, says Pasteris. She and Brigitte Wopenka, also of Washington University, presented Raman spectra from non-biogenic carbon material at the GSA "that looked just like the spectra that Schopf et al produced in Nature. In other words, the spectra in Nature are the spectra that one gets from disordered carbonaceous material, whether or not it’s biogenic in origin."

Can Raman spectroscopy "really distinguish this kerogen from some other yucky carbon material? The answer is no," Pasteris told Astrobiology News.

In other words, says Pasteris, the fact that Schopf was able to match up the spectral signatures of certain carbon compounds with visible shapes within his samples doesn’t prove their biological origins, because the spectral signatures themselves were inconclusive.

Fossil claims, Brasier writes in his email, "are better tested by falsification rather than by justification. To that end, our research group urges that future studies (older than, say, 3.0 billion years) should explore the following null hypothesis: that very ancient/alien microfossil-like structures … should not be accepted as of biological origin until all possibilities of their non-biologic origin have been tested and can be falsified."

The debate does highlight a need for clearer standards of evidence for matters paleontological and astrobiological. Ironically, despite more data, improved techniques, and increased focus on the field, scientists are finding that one of the most difficult questions to answer is also one of the most basic: "How do we really know?"