The Origins of Life
To the Editor:
In a letter to Joseph Hooker in 1863, Charles Darwin wrote: “It is mere rubbish, thinking of the origin of life; one might as well think of the origin of matter.” Although we physicists now routinely discuss the latter topic, Darwin’s pessimism 150 years ago is perhaps understandable, coming as it did before most of the edifice of modern physics was built, and before more than a century of growing knowledge about chemistry and biology would bring us closer to the brink of understanding life’s origins. But it is harder to understand the purpose of David Berlinski’s ultimately facile rumination, “On the Origins of Life” [February].
At first I thought the article was a reasonably cogent discussion of the history of how science has shed light on this seminal but still puzzling event in the history of the earth. But as I read on, it seemed that Mr. Berlinski’s purpose was not to explore the edge of knowledge but rather to expound the completely trivial fact that the edge of knowledge is fuzzy. Ultimately it became clear that his article was designed to convey the hackneyed and intellectually lazy argument that because there are things that are truly puzzling and paradoxical about nature, one should be prepared to give up trying to devise any natural explanation for them.
I found the errors and omissions in Mr. Berlinski’s in-depth discussion of molecular biology particularly telling. He harps on the red herring provided by the famous Miller-Urey experiment—which purported to demonstrate that amino acids could form naturally in the pre-biotic atmosphere—pointing out that the actual pre-biotic atmosphere was nothing like the “primordial soup” that Miller and Urey mixed up in their beakers, and lacked the “reducing” capacity necessary for the all-important chemical reactions. Fair enough. But Mr. Berlinski ignores the wealth of new data on a wonderful reducing environment associated with thermal vents at the bottom of the ocean. There is growing evidence, genetic and otherwise, that this anaerobic environment, where sulfur could play the role currently played by oxygen in exciting electrons to release energy, is associated with those organisms closest to the root of the tree of life.
Perhaps most surprising, given Mr. Berlinski’s background in mathematics, is his innumeracy in regurgitating the tired arguments about probability and the origin of life. Not only does he make the standard probabilistic error of assuming that self-replicating molecules result from totally random interactions—the same sort of analysis would argue against the formation of any complex molecule—but in doing so, he misinterprets the orders of magnitude involved. He argues, for example, that the number of nucleotide sequences of 100 nucleotides in length (the minimum thought necessary for “demonstrated ligase activity,” the first step in self-replication) is greater than the number of atoms in the universe—making the odds in favor of self-replication rather long. Well, the number of such sequences is very large, but in fact there are far more atoms in our local region of the Milky Way galaxy.
When I recently researched a book about the history of an oxygen atom, I had to learn about advances in molecular and evolutionary biology, fields with which I had not been familiar. The remarkable discoveries being made about archaea, hyperthermophiles, and the possibility of inorganic templates for the earliest self-replicating molecules filled me with wonder, excitement, and optimism. That is what science is supposed to do. But one can also decide in advance (as Darwin did in his time and Mr. Berlinski does now) that the answers are unknowable, and choose to bury one’s head in the sand.
Lawrence M. Krauss
Case Western Reserve University
To the Editor:
David Berlinski has performed the greater part of a difficult task—guiding his readers through the wilderness of origin-of-life theory. He first leads us through the complexities of modern molecular biology, explaining the key roles of large molecules like DNA, RNA, and proteins. He describes how knowledge about their function inspired the “RNA world” theory of the origin of life (by which life came into being on the basis of RNA replication), and proceeds to demolish this theory utterly. The reader is thus taken to the very edge of the promised land, which in this case would be a description of the most likely answer to the origin-of-life problem, and one that meets the constraints of “the model for what science should be”—which, as Mr. Berlinski puts it, requires that the laws of a system’s development be unique and stable.
But having taken us this far, Mr. Berlinski devotes only three paragraphs to the scientific alternative most likely to provide that answer, often called “metabolism first,” which holds that life began with an interactive network of small molecules that used an available energy source to maintain its identity and promote its further evolution. He notes that this view “may well be right,” but cautions that “there is as yet no evidence that it is true” and suggests that the answer may ultimately lie beyond the powers of scientific inquiry. At this point he and I part company, for I believe that the answer does lie in the idea of “metabolism first,” and is in fact close at hand.
Before I elaborate on this, I should point out that the case against an “RNA world” is even stronger than the one Mr. Berlinski presents. Not only were cytosine and ribose unlikely to have been present in any quantity on the early earth, but the same can also be said of adenine and guanine. Moreover, no adequate explanation of the manner in which these parts (and others) could connect together spontaneously to form RNA has ever been presented. All of the celebrated triumphs of “prebiotic synthesis” were actually accomplished through the active intervention of the experimenter, as Mr. Berlinski notes about two such cases.
An even more serious difficulty exists, one I call the termination problem. Imagine a series of units, each of which contains a plug and a socket. If we arrange circumstances so that they will connect together, they have no alternative but to join head-to-tail to form a linear array. But if we include a large excess of units that have only a plug or a socket, these units can connect on one side but then terminate any growing chain. This problem applies not only to DNA, RNA, and protein but also to the many RNA surrogates that have been put forward as additional examples of molecules that could have copied themselves, thereby initiating life.
This problem can be avoided if we assume—to return to the idea of “metabolism first”—that life began with small molecules (monomers) rather than, as the “RNA world” theory has it, with the entities produced by connecting them (polymers like DNA, RNA, and protein). Monomer mixtures can carry heredity, not as linear text as in DNA but by the presence of certain substances and the absence of others. This form of information storage, which has been called a “compositional genome,” could reproduce simply by splitting into two parts. Thus, I prefer the term “monomer world” to “metabolism first,” as the latter implies a lack of genetic ability.
For an analogy, imagine my wife gives me a shopping list (linear text) for the supermarket. If I come home merely with a copy of the list, she will not be pleased. If I return with an assortment of produce and canned goods (a compositional genome), she can tell by inspection that I have brought the desired items. The information is provided by the objects themselves rather than by the list.
Monomers can also show substantial catalytic ability, though less than the elaborate proteins and ribozymes that were produced by extended evolution. A key requirement for monomer life is an external energy source that interacts with a chemical mixture in a way that both releases the energy and organizes the mixture. No “smoking gun” experiment has as yet been devised that demonstrates the operation of such a system, but much preliminary work has been carried out. Suitable individual reactions have been explored, monomer catalysis has been amply demonstrated, and the natural occurrence of compartments (or other barriers) that could separate an evolving chemical system from its environment has been documented. I summarize the evidence for all this in a forthcoming paper.
In my opinion, the principal barrier to the demonstration of such a system is not technical but lies in the sociology of science. The intellectual elegance of the “RNA world” solution has diverted scientific interest and funding down a dead-end street. If the same effort and ingenuity were applied to the experimental exploration of “monomer world” theories, I believe that a satisfying laboratory demonstration of the early steps in the origin of life could be achieved.
New York University
New York City
To the Editor:
Around the time I read David Berlinski’s summary of the failures of origin-of-life research, I saw a report in the Economist taking up similar issues. After quoting the insights of the researcher Charles Cockell, who argues that early impact craters full of hot water were “ideal places for life to get going,” the anonymous author notes that “the biggest irony of all . . . might be that the conditions once thought a near-insuperable obstacle to the emergence of life on earth may actually have enabled it to come about.” Well, yes, that would be handy. If only sinking businesses and failing states could pull off stunts like that.
Meanwhile, on February 15, the American Association for the Advancement of Science published a short feature quoting another researcher, David Deamer, who enterprisingly developed his own primordial soup. He dumped a can of the stuff into a hot volcanic pool and found that most of the material simply disappeared, uninspired, into the clay lining of the pond. He concluded that such pools may be unlikely spots for “the first assembly of life’s little bits.”
Actually, if the recipe were really “just add water and boil,” we should see new life forms springing from chemical soups all the time. But current theory insists on at most a few common ancestors. So there must have been some magic in that old soup cauldron. Or are we on the wrong track?
Current origin-of-life research seems fundamentally incoherent. People who are convinced that life originated by an unlikely accident also believe that they can find out exactly how it happened. Origin-of-life researchers also sometimes denounce intelligent-design theory. But if the universe started with a large input of information, whether provided by a cosmic mind or an inherent principle of organization, we can safely postulate that the origin of life was not an accident—nor was the human mind that now studies them.
To the Editor:
As far back as 1897, the American biologist E.B. Wilson wrote in his textbook, The Cell in Development and Inheritance: “The study of the cell has, on the whole, seemed to widen rather than to narrow the enormous gap that separates even the lowest forms of life from the inorganic world.” I gather from David Berlinski’s characteristically brilliant article that even as late as 2006, Wilson’s assertion requires no emendation.
Earlier than Wilson, the 19th-century founder of positivism, Auguste Comte, defined the problem this way: the change of order that takes place when we pass from the inorganic to the organic is—in plain, open fact—the passage from an order in which the parts precondition the whole to an order in which the whole shapes the parts and, in a sense, precedes them, as if the parts were there only in view of the whole.
As a positivist, Comte was rightly unnerved by that ineluctable reality, for it conjures up the specter of final causality, with its implication that the mental precedes the material. But at least he noted the problem of teleology raised by the organic world instead of just wishing it away, as Francis Bacon did in his manifesto, On the Proficience and Advancement of Learning (1605), in which he compares final causes to a lamprey clinging to the side of a ship and impeding its progress. As the great French medievalist Etienne Gilson says in From Aristotle to Darwin and Back Again (1971), “The important thing is to know whether or not final causality expresses a fact given in nature, for if we object to final causality as an explanation, it still remains as a fact to be explained.”
With Mr. Berlinski, I do not deny that the invocation of final causality has gummed up scientific methodology down through the ages (Bacon’s point). But the converse must also hold true: maybe science cannot, from the outset, get at the whole of reality. Again, Gilson gets it exactly right: “If the scientist refuses to include final causality in his interpretation of nature, all is in order; his interpretation of nature will be incomplete, not false. On the contrary, if he denies that there is final causality in nature, he is being arbitrary. To hold final causality to be beyond science is one thing; to put it beyond nature is something completely different. In the name of what scientific principle could one exclude from a description of reality an aspect of nature so evident?”
Edward T. Oakes, S.J.
University of St. Mary of the Lake/Mundelein Seminary
David Berlinski writes:
The physicist Lawrence M. Krauss has become well known for the unflagging zeal with which he has defended Darwin’s theory of evolution from its critics. With admirable gusto, he has recently undertaken to wave the crutch of his concerns toward his colleagues in physics as well. String theory, he argues in Hiding in the Mirror (2005), seems to have gone nowhere, and to have gotten there rapidly. Inasmuch as the same can be said of his letter, Mr. Krauss appears to have achieved postmodern status as a perfectly self-referential critic.
It was my intention, he asserts, to “expound the completely trivial fact that the edge of knowledge is fuzzy.” Edges by their nature cannot be fuzzy; and “the edge of knowledge” suggests an inscription on a medieval map. Instead of the edge of knowledge, shall we say current research, and instead of fuzzy, inconclusive? With these corrections in place, Mr. Krauss’s trivial fact is rather less trivial than he supposes. The scientific establishment, after all, has expended a great deal of energy denying it, especially when funding is at stake.
Having made his point once, Professor Krauss makes it again. My article, he writes, “was designed to convey the hackneyed and intellectually lazy argument that because there are things that are truly puzzling and paradoxical about nature, one should be prepared to give up trying to devise any natural explanation for them.” Not at all. I am not one of those who, for example, give up on string theory at the first suggestion of intellectual difficulty. I propose to give up only when the getting out is good. In my article on the origins of life, I hesitated conspicuously.
But to specifics. Mr. Krauss concludes that my observations on the Miller-Urey experiment are “fair enough,” but he objects that I ignored “the wealth of new data on a wonderful reducing environment associated with [hydro]thermal vents at the bottom of the sea.” Those hydrothermal vents—now divided into the superbly named “black smokers,” where hot volcanic material containing metal sulphides gushes into cold sea water, creating a characteristic black spume, and “cold seepers,” whose etymology is perhaps best left unexplored—made their first appearance as origin-of-life contenders in a paper published in Science in 1979. Thereafter, the German organic chemist and patent attorney Günter Wächtershäuser promoted them to a still more prominent status by imagining a rich series of inorganic catalytic reactions that might have taken place in such environments. A number of distinct (and immensely interesting) chemical ideas were put in play, the most notable involving iron pyrite acting to promote cyclic chemical reactions among carbon, oxygen, and hydrogen.
But experiments did not produce a significant yield of crucial biological molecules, and critics (like Gerald Joyce) have complained that some of the experiments themselves were carried out under unrealistic laboratory conditions. Matthew Levy and Stanley Miller have argued in addition that, although “a high-temperature origin of life may be possible,” it “cannot involve adenine, uracil, guanine, or cytosine”—because, at such temperatures, these molecules are notoriously unstable.
In brief, Mr. Krauss’s “iron-sulphur world” is not new; nor is it “wonderful” (although it is certainly interesting); nor, most of all, is it relevant to the scenario of an “RNA world” to which my own article was largely devoted. The two scenarios, as everyone familiar with the field understands, are in conflict, the “iron-sulphur world” being a natural entryway into the “metabolism first” theories that I mentioned briefly in the conclusion of my article (and shall return to in my comments on Robert Shapiro’s letter).
It is nonetheless a fact that the “RNA world” scenario, which the “iron-sulphur world” is intended to displace, is still the dominant system of thought in origin-of-life research, still the most powerful intellectual structure in contention, and still the locus of the greatest investment of time and money. I did not discuss the “iron-sulphur world” because, in my judgment, it does not possess the same degree of intrinsic importance as its chief rival, either historically or scientifically.
There remains the charge of my “innumeracy,” a condition I allegedly attained while “regurgitating the tired arguments about probability and the origin of life.” Mr. Krauss’s premise is that I am mistaken in assuming that self-replicating molecules must have arisen from totally random interactions, for, were I correct in this, the same argument would militate against the formation of any complex molecule. Here he confuses doubts about the applicability of the theory of probability with an error in the theory itself.
What Mr. Krauss takes to be a mistake is simply a fact. When nucleotides are connected in a sugar-phosphate chain, they form a polynucleotide. Polymerization is the chemical activity involved in forming such chains. But a problem arises when questions are asked about the specific sequences of nitrogenous bases required to carry out replication (or any other biological activity). It is a problem because, while polynucleotides are formed by means of ordinary laws of covalent chemical association (given an energy source), polymerization itself is not sequence-specific. Like flags mounted on stalks, the nitrogenous bases are fixed to their sugar-phosphate backbone, but the order in which they are fixed is free.
The standard (and only) accounts of this in the various “RNA world” scenarios begin with what Leslie Orgel and Gerald Joyce call “a random pool of nucleotides.” For a self-replicating polynucleotide to be obtained from this pool, some process of polymerization must have taken place. But polymerization is sequence-specific only in the context of the laboratory or the living cell. Since neither was available in any pre- biotic era, there remains chance, and only chance, as a guiding force.
Hence the perfect relevance of probabilistic calculations. Those “tired arguments”—which by the way are not mine but have been made by Gustaf Arrhenius as well as by Orgel and Joyce—still seem to me remarkably vigorous, and at least twice as strong as the Parthenon.
Finally, I did not argue that the number of nucleotide sequences that are 100 nucleotides in length is “greater than the number of atoms in the universe.” I observed that the number of such sequences is precisely 4 to the 100th (or roughly 10 to the 60th) power, and then illustrated the magnitude of this number by comparing it with the number of atoms in the universe or the time in seconds that has elapsed since the Big Bang.
My observation was mathematically correct—obviously so. Scrupling, instead with my illustration, Krauss could, for all I know be right, although he provides no evidence or argument. Estimates in the literature tend to vary by a great many orders of magnitude; no one quite knows, for example, how many atoms to assign to the unobserved portions of the universe. But those atoms could disappear from my argument with no ill effect. Four raised to the one-hundredth power would still remain a dauntingly large number. It is embarrassing to have to point such things out.
Observing with satisfaction the various noses I punched in my article, Robert Shapiro suggests that I should have taken a shot at chain termination as well—that is, the tendency of growing polymers to branch off absurdly before achieving any form of biological usefulness. I am sure he is right about this, and he is right again in his comments about the sociology of science.
Like an immense ocean liner, big science achieves its momentum at the expense of flexibility; no matter the theory, correction and readjustment are very slow in coming, when they come at all. String theory and the “RNA world” entered the scientific imagination at roughly the same moment and have followed a similar trajectory, in which expectations have steadily outrun accomplishments. If this were the whole of the story, judgment would be easy. But both theories also have outstanding accomplishments to their credit, so that even their critics can feel the allure of the common counsel not to give up, at least not just yet.
As for “metabolism first” theories, they do indeed represent an alternative to the “RNA world” scenario. But as I have already indicated, my aim was not to survey the entire field but to construct a responsible and coherent narrative, one that describes the shape of research as it is, not as it may be.
My view of the theories that have engaged Mr. Shapiro’s interest is respectful but wary. The metabolism-first world is the work of organic chemists, reflecting a decision to establish, as the decisive origin-of-life event, the advent of certain self-sustaining or autocatalytic cycles. The reverse citric-acid cycle plays a prominent role in these deliberations, conjecturally preceded by a primitive citrate cycle that takes place within an iron-sulphur world and is catalyzed not by biological enzymes but by metal ions.
Any chemical cycle consists of a linked series of reaction steps; an autocatalytic cycle is one whose products feed back into the cycle, thus keeping it going. But even assuming the existence of something like a primitive citrate cycle driven by non-biological catalysts, no one has yet demonstrated that the quite different conditions required for conversion to a true autocatalytic system could have obtained in the pre-biotic era. The only true autocatalytic cycle remains the formose reaction, and for reasons that I discussed at length in my article, that reaction does not seem to be going anywhere.
The idea of a compositional genome, to which Mr. Shapiro also draws our attention, is largely the work of Doron Lancet’s research group at the Weizmann Institute of Science. It is meant to serve, as Lancet has said, as a “bridge between the ‘genome first’ and the ‘metabolism-first’ paradigms.” In a very broad sense, the work of Lancet and his colleagues is continuous with the work of Mannfred Eigen, Freeman Dyson, and Stuart Kaufmann in that it seeks to create a theory in which order can be seen to appear out of chaos.
The chaos in this case is the froth of chemicals sloshing around somewhere in the pre-biotic era. No particular molecule within the froth is capable of self-replication. But might the ensemble itself be capable of this, at least to the extent that, after many random chemical interactions, one will have an ensemble closely matching the compositional characteristics of the original? This is the Lancet group’s thesis, one that has been advanced by means both of a few differential equations and of a fabulous series of computer simulations.
The result is called the Amphiphilic Graded Autocatalysis Replication Model, or A-Gard, and I must confess it seems to me to reflect an inflation of trivialities. A loaf of bread, when divided, gives rise to two smaller loaves that share the compositional structure of the original. Of course they do; they are made of bread, after all. I suppose that one could call this a form of heredity, but then any division of any material structure would embody a form of heredity as well.
To me, the difference between the behavior of various chemical ensembles studied largely in computer simulations and the coded chemistry characteristic of living systems still seems absolute. Lancet’s enterprise is an attempt to make the Darwinian principles of random variation and natural selection do useful work in the pre-biotic era, where plainly they do not belong, and in my view it thus represents a step in the wrong direction.
Denyse O’Leary mentions David Deamer, who has been an important figure in origins-of-life research. As she says, reports suggest that his most recent experiments have been a flop. “The results are surprising,” Deamer himself has remarked, “and in some ways disappointing. It seems that hot acidic waters containing clay do not provide the right conditions for chemicals to assemble themselves into ‘pioneer organisms.’”
But I differ with Denyse O’Leary’s statement that there is something “incoherent” about current origins-of-life research. In many respects, the study of the earth’s pre-biotic and early biological eras are stronger, better informed, and more penetrating than research devoted to the near-history of life. We understand the issues of pre-biotic chemistry more clearly than we understand the appearance of echo-location in bats or the emergence of the mammalian visual system. The truth of the matter still remains beyond our grasp, but organic chemists and molecular biologists have given us a splendid and inspiring example of the sheer power of their methods. What we do not now know, we at least know better than we did.
I agree with the spirit of Edward T. Oakes’s remarks, and in one respect with their substance as well. The study of the modern cell has been an exercise in astonishment. The thing is complex beyond belief, and molecular biologists, as anyone reading the journals will attest, have only begun to explore that complexity. For the moment, we lack the proper tools even to offer a precise and reasonably complete account of the cell’s biochemistry and biology; a full, satisfying, and profound understanding of its nature lies far in the future.
On the other hand, I do not think that a revival of the Aristotelian categories is apt to occur any time soon. A willful blindness to the plain facts of teleology in the universe? Perhaps. Biology, even on the molecular and biochemical level, is surely pervaded by teleological thinking. The activities of the cell are purposeful, and so too are the activities of every system made from cells. No one for a moment doubts that the purpose of the heart is to pump blood. These similes—the heart is like a pump—may in the end prove ineliminable. Nonetheless, they strike molecular biologists and biochemists as impediments to thought.
I would like to suggest to Father Oakes a somewhat different perspective. Let us drop Aristotle for the moment. In the standard view of the sciences, explaining the universe is an activity best conceptualized in terms of a descending order: from mind, to life, to matter. (That indeed is how I am proceeding in this series of articles, which began in November 2004 with “On the Origins of the Mind.”) To put this in terms of scientific disciplines, psychology vanishes into biology; biology vanishes into molecular biology and then into biochemistry; further down there is organic chemistry; then come physics and, finally, mathematics. But what if the arrows pointing downward are reversed, so that, in the end, mathematics and with it our entire picture of the physical universe turns out to be crucially contingent on psychology—i.e., on mind?
Readers of Commentary eager for my third and final essay in the series may take these remarks as a preview of coming attractions.