Contemporary Work On The Origin Of Life : 13.7: Cosmos And Culture Work on the origin of life is rapidly entering a phase of significant scientific progress.
NPR logo Contemporary Work On The Origin Of Life

Contemporary Work On The Origin Of Life

Work on the origin of life is rapidly entering a phase of significant scientific progress.  I want to discuss at least my view of its contemporary status after a brief historical summary.

I begin with this: It is astonishing, but true, that molecular reproduction in the form of collectively autocatalytic sets of either DNA or RNA or peptides have been made. The most complex, created by Reza Ghadiri at Scripps and his former post-doctoral fellow, Gonen Askanazy, now at Ben Gurion University, consists in nine polypeptides, each 32 amino acids long, that mutually catalyze one another’s formation from fragments of each of these nine polypeptides.  Critically, these results demonstrate conclusively that molecular reproduction need not be based on template replicating DNA, RNA.

Equally importantly, these collectively autocatalytic sets are based on ligation reactions, in which a 32 long amino acid sequence forms an alpha helix that folds onto itself, but because of this, can bind a 15 amino acid sequence and a 17 sequence that are fragments of another of the nine polypeptides, and themselves form alpha helices. The 32 amino acid sequence then binds and ligates the two fragments that then constitute a second copy of one of the other nine 32 long polypeptides.  Here, no 32 long polypeptide copies itself, the set as a whole achieves catalytic “closure,” that is, each reaction that must be catalyzed is catalyzed by a member of the collectively autocatalytic set.

More recently Gerald Joyce at Scripps has reported experiments in which two RNA sequences form collectively autocatalytic sets, each catalyzing the ligation of fragments that form the other.  
The alternative hypothesis to autocatalysis by catalysis via ligation or other reactions is the hope to form an RNA enzyme, called a ribozyme, that can act as a “polymerase” enzyme able to catalyze the template replication of any RNA, by sequential addition of nucleotides that are Watson Crick complements of the template strand.  Such a ribozyme polymerase could replicate itself, hence be a self reproducing molecule.  No such RNA molecule has been found, but serious work to evolve such an RNA sequence from a library of RNA molecules with enormous diversity is underway and may succeed. David Bartel at the Whitehead Institute is a leader in this effort.

These and other avenues discussed below suggest significant progress towards invitro formation of living systems.

Prebiotic Chemistry

The famous Miller Urey experiments in the 1950s involved electrical sparking of a chamber with a reducing atmosphere including methane, CO2, and other compounds, with water in the system to model an “ocean”. The astonishing results were the formation of a modest variety of biologically relevant amino acids.  This results initiated many years of prebiotic chemistry experiments. In summary, amino acids are easy to make, nucleotides can be made with considerable difficulty, and other biologically important small molecules can be synthesized “abiotically”.  The general conclusion is that the “primitive soup” could have been formed on earth.

Conversely, the prebiotic soup of small organic molecules may have arrived from space.  Some meteorites are rich in organic compounds.  The famous Murchison meteorite, which fell in Australia about 40 years ago has recently been analyzed and shown to have at least 14,000 different organic molecules and the capacity to form millions from these.  Infall of such meteorites may be the only or main source of small organic molecule diversity on the early earth.

Early Efforts At Enzyme Free Template Replication of RNA Molecules

RNA molecules form double stranded helices, just as do DNA molecules, by Watson Crick pairing of nucleotides on the two complementary strands of the helix. Leslie Orgel, late of the Salk Institute, spent many years of his life attempting to get a single stranded RNA to line up the four RNA nucleotides, A, U, C, and G, from a solution of free single nucleotides, get the template to ligate these nucleotides by proper 3’-5’ phosophodiester bonds to form a second complementary strand, then melt the two strands, and cycle again. This would have achieved molecular reproduction without a polymerase enzyme.

So far, Orgel type experiments have not worked for a variety of chemical reasons, although they might yet work.


The cell membrane is a bilipid layer.  It is wonderful that the analogue of the cell membrane is simple to form, even from molecules taken from the Murchison meteorite. If cholesterol, a lipid, is dissolved in water, it forms bilipid membranes that close to form hollow spheres called liposomes.  One line of thought is that liposomes were an early self replicating structure in themselves.  L. Luisi has formed self reproducing liposomes.

More, any hope of forming autocatalytic sets of polymers such as the results of Ghadiri and Askanazy, on the primitive earth would seem to require some means to sequester the reactants such that they did not diffuse out of reaction range with one another.  
Network Catalysis In Reaction Cycles.

Some chemical reaction cycles produce an extra copy of one of the reactant organic molecules. The TCA cycle in metabolism is an example, where if run in reverse, it produces an extra copy of one of the reactants in the multistep reaction cycle.  Some workers, notably Harold Morowitz, believe that such “network catalysis” plays an essential role in the origin of life.    
A new field, Systems Chemistry, is forming, where it will soon be possible to analyze large reaction networks of organic molecules and test computationally or experimentally how many such network catalytic cycles exist and how they may interact. Work on organic reaction networks, as above, is part of a larger effort to understand how metabolism among small molecules forming the building blocks, amino acids and nucleotides and lipids, may have occurred in the origin of life.  Obviously, the advent of catalysts to speed some or many of the reactions in such a reaction network can and has occurred, for in contemporary metabolism each reaction step has a specific protein catalyst, or enzyme, such that metabolism now forms a connected network of catalyzed reactions.

Theoretical Frameworks

Three alternative but in many ways complementary theories were put forth in 1971.  Manfred Eigen put forth the theory of hypercycles. Eigen supposed a set of replicating RNA pairs of complements, each itself able to replicate as in the Orgel type experiments noted above. Then Eigen supposed that each of N such  pairs of + and - complementary RNA strands, could “help” the next one of the N + and - strands to replicate, around a cycle among the N RNA replicating pairs.  This “hypercycle model” has been studied intensively, see “The Hypercycle” by Eigen and Peter Schuster.

In the same year, T Ganti put forth the “Chemotron” model, which is the most sophisticated early model. It contains all three components above, a replicating molecule, presumably a complementary pair of RNA or RNA like strands that template replicate as in the Orgel experiments above. In addition, Ganti assumes a metabolism to supply the nucleotides for these reactions in an open far from chemical equilibrium system, and he assumes a bounding membrane to enclose these components.

The third theoretical framework was put forth by myself in the same year, and subsequently in 1986 and is in all my books, “The Origins of Order”, “At Home in the Universe”, “Investigations” and “Reinventing the Sacred”.

The core idea in my own theory of the emergence of collectively autocatalytic sets is simple and involves what is called a phase transition in a “random graph”. Consider a thousand buttons on the floor and a spool of red thread. Break off pieces of the red thread and tie a randomly chosen pair of buttons together by the thread. Iterate this process, tying randomly chosen pairs of buttons together with red thread. Every now and then, lift a button and see how many buttons you lift with it.  Such a cluster of buttons is called a “connected component” in a “random graph”, where a graph is just a set of nodes connected by lines.

As shown by Erdos and Renyi in 1959 and 1960, as the ratio of threads to buttons increases, at first there are small clusters of buttons, then middle size clusters of buttons, then a wonderful thing happens. Addition of a few more randomly tied pairs of buttons ties all or most of the middle size clusters into one giant cluster or giant component of the graph.

Mathematically this transition is, like ice to water, a “first order” phase transition. For a large number of buttons, or nodes, as the number of connecting lines or edges increases, the system rather suddenly jumps from disconnected small clusters to a giant connected cluster.

With this phase transition in mind, my idea was simple. Consider a set of linear polymers, made of two monomers, A and B.  Consider all monomers and polymers up to length M.  Consider only cleavage and ligation reactions among these, that is AB and BA can ligate to form ABBA, or ABBA can cleave to form A + BBA.

It is easy to show, and intuitively obvious, that as the number of molecules in the system grows, the number of reactions among them grows even more rapidly.  For example the longest polymer, length M, can be made by ligating smaller fragments to form each of its  M-1 bonds, hence this M length polymer can be formed in M - 1 ways.  It is easy to show that the ratio of reactions to molecules is (M - 2), thus the ratio of reactions to polymers increases “linearly” with the length of the longest polymer.  
The final step is this theory is simple. We need to know if any of the polymers can catalyze any of the reactions in the system of reactions.  A simplest hypothesis is that any molecule has a fixed probability, P, of catalyzing any reaction. Then it is easy to prove that as the diversity of molecules increases, more and more reactions are catalyzed, hence the reactions take place rapidly.  Eventually, so many reactions are catalyzed that a giant catalyzed reaction network forms which is collectively autocatalytic.

In short, on this view, the emergence of collective autocatalysis is an expected phase transition among a set of molecules that can undergo reactions and are candidates to catalyze those very reactions.  The emergence of collective autocatalysis is “expected”.  
These ideas have been shown to work in computer experiments in 1986, by Farmer, Packard and Kauffman. More such collectively autocatalytic sets can evolve to novel sets so evolution can occur.

More recently, W. Hordijk and M. Steel have proven analytically that the above phase transition occurs and that the expected number of reactions a polymer needs to catalyze to achieve autocatalytic sets is on the order of 1 to 2, which these authors consider reasonable for either RNA ribozymes, polypeptides, or both together.

In “The Origins of Order” I describe the emergence of a connected catalyzed set of metabolic reactions to support the autocatalytic set.

Experimentally, it is now possible to generate large libraries of stochastic DNA, RNA, and peptides or polypeptides. T. LaBean, and more recently, L. Luisi, have shown that stochastic polypeptides fold quite readily.  In a field called combinatorial chemistry, one subfield is called “phage display”. Here viruses that infect bacteria, called phage, are engineered such that each virus has on its surface coat a unique random peptide. Experimental work shows that about one random peptide in a million can bind to a given molecular target.  Such proteins are called peptide aptamers and are used to find drug candidates.   
Using such libraries, it should soon be possible to test whether collectively autocatalytic sets arise as diversity of the molecular system is increased.  
Three final but major points:

  1. Obviously, it is necessary to bring molecular reproduction, metabolism and something like a lipid bounding membrane and reproduction of a protocell together.
  2. The systems considered above are what biochemists call “exergonic”, they are fully spontaneous processes. Real cells link non-spontaneous processes, called endergonic, which do not occur without the addition of free energy, to spontaneous processes which supply that free energy.  This must be brought into the picture, along with the performance of thermodynamic work cycles that link spontaneous and non-spontaneous processes like an engine. Real cells do work cycles.
  3. Third, and massively important, if early life started as sketched above, a “Darwinian transition” to cells that use DNA, RNA and code for proteins via a genetic code must take place at some point. The capacity of such DNA,RNA, protein systems to evolve is almost certainly much greater than the protocells hoped for above.