If DNA is the Word, then the Word was not in the beginning. So says Freeman Dyson. DNA or RNA was only the second beginning, hooking into a pre-existing protein life like a virus parasite.

Extending Lynn Margulis’ (1981) picture of the eukaryotic cell as the result of cooperation (symbiosis) of two or three more primitive (prokaryotic) cells further back in time, Dyson sees even the nucleic acids as only later adjuncts, though necessary for subsequent evolution, as it turned out. However, if this very early symbiosis had not occurred, other lines of development, quite inconceivable to us, might have resulted.

What happened in the first beginning, he does not say, except to indicate that metabolism began as chemical cycles involving proteins and energy production and consumption. And he adds a third component of the earliest life, besides replication (DNA or RNA) and metabolism (protein), (which he calls software and hardware, respectively); the third component is spatial segregation of the initial living matter into droplets of “coacervate” (a kind of a liquid precipitate from the main aqueous phase, like the oily droplets in an emulsion; this concept was used by Oparin in his book “The Origin of Life”). The coacervate droplets were separated from the surrounding medium by a primitive membrane (somewhat like soap films stabilizing an oil-in-water emulsion) and became proto-cells, even before the metabolism inside became fully functional. The separation is necessary, because the intricate mechanisms of life can develop and stabilize with sufficient intensity to become self-sustaining in only limited regions of space, and they must be kept from being constantly diluted by the lifeless medium in which they are suspended.

I will try to speculate now about that initial push, the real first beginning. This is a bit audacious, because little is known about it. James Lovelock, the originator of the Gaia hypothesis, wrote that to him the origin of life is “ineffable”. But it is not tabu territory to speculate about, which many have done. Our very real fascination with it will not permit us to stay away.

Many years ago, Lecomte De Nouy “proved” that life could not have arisen by chance or accident, because it is just too improbable to have all the right atoms to come together by chance in the right numbers and configurations. He calculated the probability of this n-body collision, and concluded that the time required for this to happen by chance would be far longer than the known age of the universe. Ergo, he reasoned, God must have created life.

To me, that argument is fallacious, because no one has ever assumed such an n-body collision – the coming together would be gradual and stepwise, from building block to bigger building block, as in many ordinary chemical reactions. Moreover, he totally ignored inter-atomic and inter-molecular forces of attraction (valence or Van der Waals), which make atoms “stick together” in collisions instead of bouncing off elastically. Atoms are not just a bunch of billiard balls. The thermodynamic equation F = H – ST describes the situation. (F is free energy, H is total energy, S is entropy, T is absolute temperature.) F is a measure of the tendency of macromolecules to build up, H is the chemical bonding attraction which helps it along, and S is the tendency to disorder, i.e. the tendency for the macromolecules to fly apart again, which they do with increasing alacrity as the temperature goes up. So the build-up is governed by these two opposing tendencies, to build up and tear down. The entropic tendency to tear down large improbable structures like proteins is admittedly very high, and therefore the probability of their formation and preservation is very low – but not quite as low by far as Lecomte De Nouy calculated. The universe has had enough time to form proteins spontaneously, but probably not to preserve them for continuous operation, without further processes helping out.

Thermodynamics is all about equilibrium; but living systems are Prigoginian structures – far from equilibrium, open to through-flow of matter and energy, and continuously self-regenerating through intricate cycles. These are called dissipative structures, and they build up high negentropy inside by exporting entropy outside. (This is why even primitive proto-cells need a membrane – to distinguish the inside from the outside.) Not all dissipative structures are “alive” (whatever that means – I have not even defined “life”); some are just self-perpetuating chemical reactions like the Belousov Zhabotinsky reaction maintained by Prigogine in his “Brusselator”, or the patterns one gets by heating shallow layers of water. But of course, the most interesting dissipative structures are alive, and Prigoginian theory is interesting mainly because of this connection.

The main feature of Prigoginian dissipative structures is pattern maintenance; while matter and energy flow through quite freely and sometimes quite rapidly, the pattern is preserved, in what has been called a hyper-stable condition. In the case of early life, “the pattern” refers to the metabolic energy cycles that Dyson mentions. The initial push to form the pattern would come from the energy of sunlight, lightning bolts, or volcanism on the primitive earth in a reducing atmosphere.

It is known from Miller’s experiments that, even today, we can form amino acid solutions by passing electrical discharges through a reducing mixture composed of water, methane, ammonia, and hydrogen sulfide. From the reactants to the products, the path is not very steeply up in the free-energy landscape, and the amino acids formed are fairly stable. Amino acids are even found in meteorites, so they are not that uncommon in the universe. They are less stable in today’s oxygen-rich atmosphere, but now that does not matter any more. (Carbon dioxide is the most stable form of carbon under today’s conditions.)

Amino acids are the building blocks of proteins, but they don’t hook up (polymerize) spontaneously. There is a free-energy hill to climb up to proteins. Perhaps (this is speculative) the energy transfer needs an intermediate, like the ATP-ADP-AMP system (or some other molecules with high-energy phosphate bonds). (ATP is adenosine triphosphate, ADP is adenosine diphosphate, AMP is adenosine monophosphate; they form from each other by the addition or subtraction of phosphate groups, absorbing or liberating energy in the process.) This system could be “charged” by solar energy and then help polymerize the amino acids to proteins. The proteins would eventually fall apart again (depolymerize), so we still have no stable cycle. Repeated injections of energy would still be needed. We have not yet achieved a take-off into self-sustainability.

The proteins formed would vary in their amino acid composition, the sequence of the amino acids along the polymer chain, and therefore also the conformation (the folding of the chains, overall shape, and the polar and non-polar groups displayed on the surface). There are some 20 different amino acids that compose proteins today; the assortment may have been somewhat different in the beginning, but the number was probably about the same. This would give extremely high numbers of possible combinations and permutations in the sequences (assuming they form at random), and therefore in conformations. Every time a polymerization to protein would occur, the product would be somewhat different, as in a giant lottery machine blindly seeking for a most favorable conformation. There would not yet be any RNA guiding the synthesis of specific proteins, no ribosome machinery for the assembling of amino acids – only the energy from ATP and the supply of assorted amino acids as the raw materials. The selection would be quite unguided at this point, as it needs to be to find the right combinations, which are as yet unknown. If we don’t know the combination in a padlock, all we can do is dial numbers at random, hoping eventually to unlock the treasure box. It might take eons, but we have the time. (It does not take as long as the age of the universe.)

One of these protein macromolecules some day might have the right shape for becoming an enzyme; i.e. have a “site” on its surface, a pocket of the right size and polarity, to catalyze a chemical reaction, and have at least a reasonable specificity for that one reaction only. The specificity and catalytic power might not at first be as high as that in modern enzymes of present living forms, which are very “high tech”, but be sufficient though crude tools for their time. The chemical reaction catalyzed by the proto-enzyme might be some fermentation, e.g. producing energy from hydrogen sulfide or methane, as some Archeobacteria still do today. The energy from the fermentation would then charge up the ATP system which could produce more protein. It could also produce the materials for the bilipid layer for cell membranes and supply other needs. We now have a stable sustained energy supply, no longer dependent on bolts of lightning and the like; we can produce proteins rapidly, speeding up the selection of more enzymes, and improving the efficiency of previous enzymes. New and better enzymes can catalyze new and better fermentations producing more energy. We have achieved take-off to a self-sustaining, even self-accelerating cycle.

It can feed on itself and maintain its pattern, like a true dissipative structure. It still does not “breed true” in the sense that the same enzyme may not be produced in the next polymerization, because it is unguided. Exact replication has to wait for the nucleic acids, RNA or DNA, which are not yet there. But, as Dyson says, high variability may be an advantage in the beginning, when the search for the right combinations is still on. It is only when the really excellent combinations have been found that we want to hang on to them for sure and replicate them faithfully.

In a memorable passage, Dyson talks about the boundary between life and death being very fluid at this stage, and moreover easily reversible. It would have been just as easy to originate life in repeated little jumps and revert to non-life just as repeatedly; death and resurrection as common-place recurrent events. I would compare it to the flicker in a fluorescent lamp just before it finally takes off to a steady light. (Or sometimes it doesn’t.)

Besides the detailed chemistry, what is important to me is the holistic image: a strong jet of energy from the sun keeping an unstable (but hyper-stable) structure balanced at a free-energy peak. Imagine a strong air or water jet pointing vertically upward (like the famous water-jet in Geneva), with a ping pong ball balanced on top of it; the ball keeps on falling under gravity, but the jet keeps it up, dancing forever in its precarious perch way up the free-energy mountain. That is what has happened on this planet, and the sun is still keeping our ping pong ball high up, in fact ever higher over time, by its strong whoosh of energy flux. But it can do it only because the energy transfer systems are in place, the fermentations and the ATP and the amino acid polymerization and the enzymes. (Nowadays also photosynthesis and respiration, but these came later.)

By focusing both on the macro and the micro picture, the bouncing ping pong ball and the chemical cycles, the holistic and the reductionist image, we can gain an overall insight into this event of primal Genesis. It was not really a point event, but an extended historical process. Understanding of it is not only an intellectual process, but a spiritual experience. There is a “Eureka” feeling of revelation in finally but suddenly grasping the complex aspects in their totality. This is what happened on this planet! My God!

The story is, of course not complete as told. We must proceed now to the “second origin” of Dyson, the introduction of the replicator systems. I will stick quite closely to his description here, which I found very enlightening.

At the stage of protein life as described up to this point, there would have been present in the proto-cells a certain concentration of the members of the ATP system and some of its chemical cousins. We need more chemical details here: adenosine monophosphate (AMP) is an example of a nucleotide; its chemical cousins would be other nucleotides, such as uracil, guanine, thymine, or cytosine. These too might be active in energy transfer systems in the proto-cell, because they too can add extra phosphate groups and hold them by high-energy bonds. Now the nucleotides are in the same relation to nucleic acids (RNA and DNA) as amino acids are to proteins; i.e. nucleotides can be polymerized to nucleic acids like amino acids can to proteins.

Experiments by Eigen (described by Dyson) have shown that nucleotides polymerize to nucleic acid if there is a protein present to act as an enzyme for this process. So we can visualize that in an early proto-cell of protein life, an “accident” happened: an RNA was produced by a spontaneous polymerization. (“An Eigen reaction 3 billion years before Eigen did it”, says Dyson.) An Eigen reaction produces an RNA with a random assortment of nucleotides (in the absence of a pre-existing RNA or DNA as a template). However, even the random RNA would have the capacity to replicate itself with high fidelity, by essentially the same base-pairing mechanisms that are now well known in molecular genetics. So the product of the “accident” would be like a highly infectious virus affecting the proto-cell and making it “sick”. (Later it was also found that RNA can act as an enzyme as well as a template; this configuation is called a ribozyme.)

Like a virus, the RNA can replicate itself only inside a living cell where the appropriate enzymes are available, not outside. This is still true of modern viruses, whether of the DNA kind (most of them) or the RNA kind (retro-viruses, like HIV that causes AIDS). It has long been a puzzle for biologists to decide whether bacteria or viruses came on the scene first – in fact, whether viruses were alive at all. We may now be close to an answer: what we might call “akaryotic cells” (ones without any nucleic acids) came on the scene first, viruses second, and their union produced bacteria, i.e. prokaryotes, with which the story as told in our text books usually starts.

How did the akaryotes (what I previously here called “protocells”) overcome the “sickness” introduced accidentally by the newly synthesized RNA virus? This is a case of converting a parasitic relationship (where the parasite exploits the host to the parasite’s benefit and the host’s disadvantage) gradually through reluctant tolerance to coexistence to adaptation and finally to symbiosis, a cooperative relationship in which both partners benefit. A zero-sum game becomes non-zero sum and then cooperative, a win-lose game goes to a win-win game.The two “origins” of life, the metabolizers and the replicators (proteins and nucleic acids), co-evolved not only to be complementary in their functions, but became so tightly coupled that they are now almost inconceivable without each other. “Overcoming your enemies by converting them into friends” has never seen a better example. DNA transcription into RNA, and RNA translation into protein, now requires specific protein catalysts (enzymes), and proteins cannot be formed other than by translation from RNA. “The Central Dogma” of molecular genetics dictates in which direction the genetic information flows, and there are few transgressions from this dogma.

You might wonder what happened to flexibility and the random search for perfection in enzymes and – yes – the replicators themselves. Dyson has a theory, which he calls his “toy model” because it is so oversimplified, according to which it was an advantage at first to be able to tolerate a high error rate (about 20%) in copying macromolecules, because it then requires fewer types of molecules (maybe 100) to put together a smoothly working homeostatic system. With the very low error rate now achieved with RNA and DNA replication, it would need some 10,000; it would probably never get off the ground. However, the advantage is reversed later on, and very accurate replication becomes desirable. With certain enzymatic “proof-reading” and “correcting” steps, genetic information is now copied with an error rate of about 1 in a billion. (As a professional editor, I am green with envy.)

The arrival of the replicators can be seen in several ways. In the ping pong ball metaphor, it means that another big whoosh of the water jet has put the little ball sky-high above the high-entropy background, like a rocket that almost disappears from sight. We are going to increasingly less probable structures, sucking negentropy from the surroundings very greedily as we go.

Another view is that life has come under a dictatorship of the replicators; that “selfish genes” now rule the roost, not really caring about the well-being of their temporary protein envelopes as long as they can reproduce and perpetuate themselves. (This is the view of Richard Dawkins.) Genes are potentially immortal, and by a strange reversal of roles, genes are the only ones that matter; the rest is a “mortal coil” to be shuffled off and a new one found as time goes on. Some genes are even “neutral” (convey no evolutionary advantage); these are carried along as “free riders” and have been termed “junk DNA” (the non-coding portions that cannot direct the making of proteins). Some genes even “cheat”: hook themselves onto successful genes to keep from being eliminated by natural selection if their gene products (proteins) prove to be disadvantageous.

Such “biological fascism” then leads to socio-biological theories of behaviours like altruism (in ants and in humans) purely on the basis of wanting to benefit one’s genes. (Sociobiology theory was developed by E.O. Wilson.) I help my own child in preference to my niece or nephew, because I share more genes with my child. I do not accept the rule of selfish genes, either as a natural fact or as a rule of moral conduct. Genes are the servants of the organism, not its masters. My choice of whose life to save first in a mass drowning is the choice of my brain, not my genes – and I might well choose a genetically unrelated friend, or in fact my husband whose genes are different. Realistically, I would probably just grab whoever is closest, if I could still think straight at all under the circumstances.

The facet of the gene-protein symbiosis that I prefer to think about is the close coupling. I don’t know what that means in chemical concreteness, but I am reminded of some of the authors in the “Gaia” book (Bunyard and Goldsmith) writing about the close coupling that occurred at some early point between the Earth’s biota and the atmosphere/water/soil. They insist that more than co-evolution is involved; that would be a loose coupling only. Close coupling at some point became inevitable, like with a teen-age boy and girl being put in bed together naked. As with the biota and the rest of the biosphere, so with proteins and nucleic acids, close coupling was inevitable, even though the relationship began as a virus infection.

I will not pursue the story of the origin of life any further, or I would end up writing a biology text book. Even so, the events of the “double origin” that I described may well have taken more time in Earth’s history than the rest of evolution, with its myriad different life forms. As with so many phenomena, the beginnings are the hardest.

Only one more question: was the development of life on Earth inevitable? Being such an improbable event, or series of events, I don’t think so; things could have gone otherwise, even under the favorable circumstances we had here. The experiment could still end, as it almost did in the Great Permian Extinction. But I refrain from speculating – this time – about ends. This essay is about origins only.