by Nick Lane
Elegant as it is, there are two serious problems with the ‘RNA world’. First, ribozymes are not very versatile catalysts, and even allowing for the most rudimentary catalytic efficiency, there is a big question mark over whether they could have brought a complex world into existence. To me they are rather less suitable than minerals as the original catalysts. Metals and minerals are found at the heart of many enzymes today, including iron, sulphur, manganese, copper, magnesium, and zinc. In all these cases, the enzyme reaction is catalysed by the mineral (technically, the prosthetic group), not the protein, which improves the efficiency rather than the nature of the reaction.
Second, more importantly, there is an accounting problem with energy and thermodynamics. The replication of RNA is work, and therefore requires an input of energy. The requirement for energy is constant, because RNA is not very stable, and is easily broken down. Where did this energy come from? There were plenty of sources of energy on the early earth, mooted by astrobiologists—the impact of meteorites, electrical storms, the intense heat of volcanic eruptions, or underwater hydrothermal vents, to name but a few. But how these diverse forms of energy were converted into something that life could use is rarely described—none of them is used directly, even today. Probably the most sensible suggestion, which has been in and out of favour over several decades, is the fermentation of a ‘primordial soup’, cooked up by a combination of all the various forms of energy.
The idea of a primordial soup gained experimental support in the 1950s, when Stanley Miller and Harold Urey passed electric sparks, to simulate lightning, through a mixture of gases believed to represent the earth’s early atmosphere—hydrogen, methane, and ammonia. They succeeded in producing a rich mixture of organic molecules, including some precursors of life, such as amino acids. Their ideas fell out of favour because there is no evidence that the earth’s atmosphere ever contained these gases in sufficient quantities; and organic molecules are far harder to form in the more oxidizing atmosphere now thought to have existed. But the existence of plentiful organic material on comets has brought us round full circle. Many astrobiologists, keen to link life with space, argue that the primordial soup could have been cooked up in outer space. The earth then received generous helpings in the huge asteroid bombardment that pockmarked the moon and earth for half a billion years from 4.5 to 4 billion years ago. If the soup really did exist, then perhaps life could have started out by fermenting a soup after all.
But there are several problems with fermentation as the original source of energy. First, as we have seen, fermentation stands apart from both respiration and photosynthesis, in that it does not pump protons across a membrane. This leads to a discontinuity and a problem with time. If all the fermentable organic compounds came from outer space, then the nutrient supply should have begun to run out after the great asteroid bombardment drew to a close 4 billion years ago. Life would dribble away to extinction unless it could invent photosynthesis, or some other way of producing organic molecules from the elements, before the fermentable substrates ran out. And this is where we run into a problem with time. Traces of fossil evidence suggest that life on Earth began at least 3.85 billion years ago, and that photosynthesis evolved some time between 3.5 billion and 2.7 billion years ago (although this evidence has been questioned lately). Given the discontinuity between fermentation and photosynthesis—not a single intermediary step brings us any loser to the evolution of photosynthesis—then the gap of at least several hundred million years, and perhaps a billion years, looks very awkward. With no other source of energy, could the organic molecules delivered by asteroids really nourish life for that long? It doesn’t sound very likely to me, especially given the tendency of ultraviolet radiation to break down complex organic molecules in the days before the ozone layer.
Second, the perception of fermentation as simple and primitive is wrong. It reflects our prejudice that microbes are biochemically simple, which is untrue, and dates back to the ideas of Louis Pasteur, who described fermentation as ‘life without oxygen’, implying simplicity. But Pasteur, as we have seen, admitted to being ‘completely in the dark’ about the function of fermentation, so he could hardly conclude that it was simple. Fermentation requires at least a dozen enzymes, and, as the first and so only means of providing energy, can be seen as irreducibly complex. I use this term deliberately, for it has been presented by some biochemists to argue that the evolution of life required the guiding hand of a Creator—that life is only possible following ‘intelligent design’. I disagree with this position, as any evolutionary biologist would, but the objection nonetheless must be tackled, and can present problems in some cases. In the case of fermentation, it is genuinely hard to see how all these interlinked enzymes could have evolved as a functional unit in an RNA world that was not supplied by any other form of energy. But notice that I am specifying, ‘a world not supplied by any other form of energy’. What we need is a means of generating energy that is ‘reducibly complex’. So the problem we must wrestle with is not how fermentation could have evolved without any other source of energy, but where the energy necessary for its evolution came from. If photosynthesis evolved later, and fermentation is too complex to evolve without an energy supply, we are still left with respiration as a further possibility. Could respiration have evolved on the early earth? The usual objection is that there was very little oxygen available on the primordial earth (see my book, Oxygen: The Molecule that Made the World for a discussion of this), but this is not an obstacle. Other forms of respiration use sulphate or nitrate, or even iron, instead of oxygen—and all of them pump protons across a membrane. They are therefore far closer in their basic mechanism to photosynthesis and hint at possible intermediary steps. Notice that this places the evolution of respiration before photosynthesis, as Otto Warburg suggested in 1931. So we are faced with the question: is respiration, too, ‘irreducibly complex’? I shall argue that it is not—on the contrary, it is almost an inevitable outcome of the conditions on the primordial earth—but before we think about this we need to consider a final, fatal, objection to fermentation as primitive.
This third objection relates to the properties of LUCA, the Last Universal Common Ancestor of all known life on Earth. Some very interesting data suggest that classical fermentation did not exist in LUCA; and if fermentation did not exist in LUCA, then presumably it did not exist in the earlier forms of life, dating back to the very origin of life, either. These data come from Bill Martin, whom we met in Part 1. There we considered the three domains of life—the archaea, the bacteria, and the eukaryotes. We saw that the eukaryotes were almost certainly formed by the union of an archaeon and a bacterium. If so, then the eukaryotes must have evolved relatively recently, and LUCA must have been the last ancestor common to the bacteria and the archaea. Martin employs this logic in considering the origin of fermentation. Up to a point, we can assume that any basic properties shared by bacteria and archaea, such as the universal genetic code, were inherited from this common ancestor, whereas any major differences presumably evolved later. For example, photosynthesis (to generate oxygen) is found only in the cyanobacteria, the green algae and the plants. Both the plants and algae are impostors—they rely on their chloroplasts for photosynthesis, and these are derived from cyanobacteria. Thus we can say that photosynthesis evolved in the cyanobacteria. Crucially, it is not found in any archaea at all, or in any other group of bacteria besides the cyanobacteria, so we can infer that photosynthesis evolved in the cyanobacteria alone, and that this happened after the split between the bacteria and archaea.
Returning to fermentation, let’s apply the same argument. If fermentation were the first means of generating energy, then we should find a similar pathway in both the archaea and bacteria, just as we find the universal genetic code in both—both inherited it from their common ancestor. Conversely, if fermentation only evolved later, like photosynthesis, then we would not expect to find fermentation in both the archaea and the bacteria, but only in some groups
. So what do we find? The answer is interesting, for both the archaea and bacteria do ferment, but do so by using different enzymes to catalyse the steps. Several of them are completely unrelated. Presumably, if the archaea and bacteria do not share the same enzymes for fermentation, then the classical fermentation pathway must have evolved later on, independently, in the two domains. This means that LUCA could not ferment, at least as we know it today. And if she could not ferment, then she must have got her energy from somewhere else. We are forced to draw the same conclusion for a third time—fermentation was not the primordial source of energy on Earth. Life must have started another way, and the idea of a primordial soup is wrong, or at best irrelevant.
The first cell
If proton pumping across a membrane is fundamental to life, as I have argued, then on the same basis it should be present in both bacteria and archaea. It is. Both have respiratory chains with similar components. Both use the respiratory chain to pump protons across a membrane, generating a proton-motive force. Both share an ATPase that is basically similar in its structure and function.
Although respiration is far more complex than fermentation today, when pared down to its essentials it is actually far simpler: respiration requires electron transport (basically just a redox reaction), a membrane, a proton pump, and an ATPase, whereas fermentation requires at least a dozen enzymes working in sequence. The main problem with respiration evolving early in the history of life is the need for a membrane, as Mitchell himself appreciated (he discussed it in a presentation in Moscow in 1956). Modern cell membranes are complex, and it’s hard to imagine them evolving in an RNA world. Of course, simpler alternatives exist. The problem with them is that they are largely impermeable to anything. An impermeable membrane obstructs exchange with the outside world, which in turn prevents metabolism, and so life itself. Given that a membrane does seem to have figured in LUCA’s respiration, can we infer from the modern archaea and bacteria what kind of membrane it might have been?
The answer betrays an extraordinary evolutionary divide with the most profound implications, which were elucidated by Bill Martin and Mike Russell (from the University of Glasgow) at the Royal Society of London in 2002. The membranes of bacteria and archaea are both composed of lipids, but beyond that they have very little in common. The membrane lipids of bacteria are made up of hydrophobic (oily) fatty acids bound to a hydrophilic (water-loving) head, by way of a chemical bond known as an ester bond. In contrast, the membrane lipids of archaea are built up from branching 5-carbon units called isoprenes, joined together as a polymer. The isoprene units form numerous cross-links, giving the archaeal membrane a rigidity not found in bacteria. Furthermore, the isoprene chains are bound to the hydrophilic head by a different type of chemical bond, known as an ether bond. The hydrophilic heads of both bacteria and archaea are made of glycerol phosphate—but each uses a different mirror-image form. These are no more interchangeable than the left and right hands of a pair of gloves. Lest such differences may seem modest, bear in mind that all the components of the lipid membranes are made by the cell using specific enzymes via complex biochemical pathways. As the components are different, the enzymes needed to make them are also different, and so too are the genes that code for the enzymes.
The differences in both the construction and the final composition of bacterial and archaeal membranes are so fundamental that Martin and Russell came to the conclusion that LUCA (the last universal common ancestor) could not have had a lipid membrane. Her descendents must surely have evolved lipid membranes independently later on. But if she practiced chemiosmotics, as we have seen she almost certainly did, then LUCA must equally surely have had some sort of membrane across which she could pump protons. If this membrane was not made of lipids, what else could it be made of? Martin and Russell gave a radical answer: LUCA they say, might have had an inorganic membrane, a thin, bubbly layer of iron-sulphur minerals, enclosing a microscopic cell, filled with organic molecules.
According to Martin and Russell, iron-sulphur minerals catalysed the first organic reactions, to produce sugars, amino acids and nucleotides, and eventually perhaps the ‘RNA world’ discussed earlier, in which natural selection could take over. Their Royal Society paper gives a detailed insight into the kind of reactions that might have taken place; we will confine ourselves to the energetic implications, profound enough in themselves.
Full metal jacket
The idea that iron-sulphur minerals, such as iron pyrites (fool’s gold), may have played a role in the origin of life dates back to the discovery of ‘black smokers’, three kilometres under the ocean, in the late 1970s. The black smokers are hydro-thermal vents—large, tottering black towers, superheated at the high pressures of the sea floor, billowing ‘black smoke’ into the surrounding ocean. The ‘smoke’ is composed of volcanic gases and minerals, including iron and hydrogen sulphide, which precipitate out in the surrounding waters as iron-sulphur minerals. The greatest surprise was that the smokers are full of life, despite the high temperatures and pressures, and the absolute darkness. An entire ecological community thrives there, gaining its energy directly from the hydro-thermal vents, apparently independent of the sun.1
Iron-sulphur minerals have an ability to catalyse organic reactions—as indeed they still do today in the prosthetic groups of many enzymes, such as iron-sulphur proteins. The possibility that iron-sulphur minerals might have been the midwife of life itself, catalysing the reduction of carbon dioxide to form a plethora of organic molecules in the hellish black smokers was developed by the German chemist and patent attorney Gunter Wächtershäuser, in a brilliant series of papers in the late 1980s and 1990s. One researcher exclaimed that reading them felt like a stumbling across a scientific paper that had fallen through a time warp from the end of the twenty-first century.
Wächtershäuser conceived of these first organic reactions taking place on the surface of iron-sulphur minerals. His ideas seemed to be supported by genetic trees, which suggested that the hyperthermophiles (microbes that thrive at high pressure and searing temperatures) are among the most ancient groups in both bacteria and archaea. However, this genetic evidence has been questioned recently (see de Duve, for example) and Wächtershäuser’s postulated reactions have been criticized on thermodynamic grounds. Perhaps most importantly, the black smoker story also suffers from a dilution problem. Once the precursors have reacted on the two-dimensional surface of a crystal, they dissociate and are free to diffuse away into the widest reaches of ocean. There is nothing to contain them unless they remain attached; and it is hard to envisage the fluid cycles of biochemistry evolving in a fixed position on the surface of a mineral.
Mike Russell put forward an alternative set of ideas in the late 1980s, and has been refining them ever since, most recently in collaboration with Bill Martin. Russell is less interested in the large, menacing black smokers than in more modest volcanic seepage sites. One such site is the 350 million-year-old iron pyrites deposits at Tynagh in Ireland. The minerals form huge numbers of tubular structures, about the size of pen lids, as well as bubbly deposits, which Russell postulates may have been similar to the hatchery of life itself. Such bubbles were probably formed, he says, by the mixing of two chemically different fluids: hot, reduced, alkaline waters that seeped up from deep in the crust, and the more oxidized and acidic ocean waters above, containing carbon dioxide and iron salts. Iron-sulphur minerals such as mackinawite (FeS) would have precipitated into microscopic bubbly membranes at the mixing zone.
8 Primordial cells with iron-sulphur membranes.
(a) Electron micrograph of a thin section of iron-sulphur mineral (pyrites) from Tynagh, Ireland, 360 million years old. (b) Electron micrograph of structures formed in a laboratory by injecting sodium sulphide (NaS) solution, representing hydrothermal fluid, into iron chloride (FeCl2) solution, representing the iron-rich early oceans.
This is not just speculation. Russell and his long-term collaborator, Alan Hall, have simulated it in the laborato
ry. By injecting sodium sulphide solution (representing the hydrothermal fluid oozing up from the bowels of the earth) into iron chloride solution (representing the early oceans) Russell and Hall produced a host of tiny, microscopic bubbles, bounded by iron-sulphide membranes (Figure 8). These bubbles have two remarkable traits, which make me believe that Russell and Hall are thinking along the right lines. First, the cells are naturally chemiosmotic, with the outside more acidic than the inside. This situation is similar to the Jagendorf-Uribe experiment, in which a pH difference across the membranes was enough to generate ATP. Because Russell’s cells come with a natural pH gradient, all that the cells need to do to generate ATP is to plug an ATPase through the membrane, surely orders of magnitude easier to evolve than an entire functional fermentation pathway! If the first step on the way to the origin of life required little more than an ATPase, then Racker’s prescient description of the ATPase as the ‘elementary particle of life’ may have had more substance than even he could have known.
Second, the iron-sulphur crystals in the bubbly membranes conduct electrons (as indeed do the iron-sulphur proteins that still exist in mitochondrial membranes today). The reduced fluids that well up from the mantle are rich in electrons, whereas the relatively oxidized oceans are poor in electrons, setting up a potential difference across the membrane of several hundred millivolts—quite similar to the voltage across bacterial membranes today. This voltage stimulates the flow of electrons from one compartment to the other through the membrane. What’s more, the flow of negatively charged electrons draws positively charged protons from inside, giving rise to a rudimentary proton pumping mechanism.