Kinematic Self-Replicating Machines
© 2004 Robert A. Freitas Jr. and Ralph C. Merkle. All Rights Reserved.
Robert A. Freitas Jr., Ralph C. Merkle, Kinematic Self-Replicating Machines, Landes Bioscience, Georgetown, TX, 2004.
4.1.6 Autocatalysis and Autocatalytic Networks
Autocatalysis – a process in which the product of a chemical reaction acts as a catalyst for that same reaction – is perhaps the simplest form of molecular self-replication. Small autocatalytic reactions have long been known, for example, the formation of the porphyrin ring system from formaldehyde and pyrrole in the presence of iron and peroxide, as described in 1969 by Calvin , and other purely chemical oscillating reactions (reaction-diffusion type replication) such as the well-known Belousov-Zhabotinsky reaction [1586-1590] (see Noyes’ explanation of it ), self-replicating spots and pulses [1593-1601] based on the Gray-Scott model , Turing patterns [1601-1609] inspired by Turing’s theory of chemical spatial pattern formation , and other oscillating chemical systems [1609-1613].
In 1986, von Kiedrowski  initiated the experimental field of “self-replication in chemistry” with the first demonstration of a simple chemical self-replicating system. As summarized by chemist Philip Ball : “When DNA replicates in cells, enzymes unzip the double helix and build two new strands, one base at a time, using the exposed single strand as a template. In the 1980s, Leslie Orgel of the Salk Institute in San Diego, California, found a way to copy short single strands without enzymes. Individual base pairs could assemble on a template strand and link up into a complementary strand. In general you can’t make two identical strands from one in a single step, however, because the new strand is complementary to, not identical to, the template. But in 1986, von Kiedrowski realized that you could get around this by choosing to copy a sequence whose complementary sequence is identical to the original when read backwards. In this way his team demonstrated enzyme-free replication of a six-base strand of nucleic acid by using it as the template for assembling and linking together two three-base fragments.” Thus all the basic design principles and all the basic findings were already published in 1986. The general scheme: A + B + C ==> ABC → C2 ==> 2 C, where C is a self-complementary template, was general enough that it could be implemented using a variety of different molecules. Von Kiedrowski used short nucleic acids because he thought that if an RNA world ever existed, short pieces of nucleic acids must be able to replicate in the absence of enzymes.
In 1989, Rebek [1616-1620] reported the synthesis of supermolecules that could generate copies of themselves when placed in a sea of simpler molecular parts, with each component part consisting of up to a few dozen atoms, a process known as autocatalysis. Unlike the bipartite complementarities (e.g., plus and minus) found in the double strands of nucleic acids and in the complementary surfaces of certain peptides , Rebek’s molecule exhibits self-complementarity or “templating” and thus provides a minimal system for replication. Illustrated schematically and as a molecular diagram in Figure 4.6, the sigmoid line represents the intermolecular contact between the two complementary and identical “AB” components. These units can be further broken along the jagged line into two different, yet complementary, pieces, “A” and “B”. The cycle of self-replication is strongly reminiscent of the several mechanical block replicators described in Chapter 3. In this case, molecule AB (imide amide) catalyzes its own formation from its simpler components, A (aminoadenosine amine) and B (pentafluorophenyl imide ester), and this reaction then causes the two AB units to disengage, each ready to resume replication with the next partners – a close molecular analog of the simple Penrose block model (Section 3.3).
Molecular self-replication has continued to be studied and discussed by chemists [1367-1370]. There are many variants of autocatalytic networks, such as those investigated both experimentally and theoretically by Bagley et al [1630-1632], Banzhaf et al [1633-1638], Boerlijst and Hogeweg [344, 1639], Cousins et al , Farmer et al [1628-1630], Ganti [1641-1646], Ghadiri et al [1382-1384], Hofbauer and Sigmund , Kauffman [1625-1628], Rasmussen , Rossler , Schuster , Stadler et al [1651-1656], Stassinopoulos et al [1657, 1658], Varetto , von Kiedrowski [1361, 1367, 1660], and Wachtershauser . One well-developed example of autocatalytic networks is the theory of hypercycles [1662-1670] as first postulated by German biochemist Manfred Eigen [1662, 1663]. Hypercycles are a connected network of functionally coupled, self-replicating chemical entities. Hypercycle reactions are interdependent, so no single reaction can be so successful that it drives out the other functions of a cycle – it is, in effect, a balanced ecosystem. Ghadiri  reported a self-replicating peptide in 1996, showing that a 32-amino-acid peptide, folded into an alpha-helix and having a structure based on a region of the yeast transcription factor GCN4, can autocatalyze its own synthesis by accelerating the amino-bond condensation of 15- and 17-amino-acid fragments in solution. In 2001, a chiroselective peptide replicator was also demonstrated , and Ghadiri’s laboratory has studied more complicated autocatalytic cycle networks [1382-1384]. It is unknown how complex such assembly cycles may be made, but Ghadiri has investigated a self-replicating 256-component molecular ecosystem incorporating ~32,000 possible binary interactions .
Self-ligating polynucleotide systems [1365-1367, 1373] also are known, showing that autocatalytic systems based on specific ligation reactions are possible. For example, the simple and successful self-replicating molecular system described by von Kiedrowski  uses a single-stranded DNA hexamer and its two trimer fragments, demonstrating (as with Ghadiri’s replicating peptide) a polymer that catalyzes its own formation from two fragments. Template-based enzyme-free self-replication of nucleotide analogs , RNA , or longer DNA sequences  has been shown. A polynucleotide system based on a ribozyme polymerase able sequentially to add the correct nucleotides (and thus copy itself) has also been proposed , and in 2001 Bartel et al  synthesized a ribozyme (itself 189 nucleotides in length) that can make complementary copies of RNA strands up to 14 nucleotides long, regardless of sequence, with 98.9% accuracy. Paul and Joyce  developed a self-replicating system based on a ribozyme that catalyzes the assembly of additional copies of itself through an RNA-catalyzed RNA ligation reaction – in particular, the R3C ligase ribozyme (size ~200 nucleotides or ~68,600 daltons) was redesigned so that it would ligate two substrates to generate an exact copy of itself, which then would behave in a similar manner. (The autocatalytic rate constant was 0.011 min-1 whereas the initial rate of reaction in the absence of pre-existing ribozyme was only 3.3 x 10-11 M-min-1.) In principle, such autocatalytic polynucleotide systems [1361-1367, 1373] could be combined with the elaborate DNA-based 3-dimensional structures, switches and motors produced by Nadrian Seeman’s group [1434-1437, 1440-1446], Jaeger’s tectoRNA geometrical figures [1675-1678], or the protein-based 3-dimensional “nanohedra” of Padilla et al  to create even more complex self-replicating molecular machines.
Last updated on 1 August 2005