Life out of chaos
Gustaf Arrhenius

Scripps Institution of Oceanography,
University of California San Diego, La Jolla, California 92093-0220, USA

ABSTRACT
Doctrinary overlays on the definition of life can effectively be avoided by focusing discussion on microorganisms, their vital processes and genetic pedigree. To reach beyond these present and highly advanced forms of life and to inquire about its origin it is necessary to consider the requirements imposed by the environment. These requirements include geophysically and geochemically acceptable conjectures for the generation of source compounds, their concentration from dilute solution and their selective combination into functional biomolecules. For vital function these macromolecules require programming in the form of specific sequence motifs. This critical programming constitutes the scientifically least understood process in the origin of life. Once this stage has been surpassed the laws of Darwinian evolution can operate in ways that are understood and experimentally demonstrated.

THE DEFINITION
In the conference from which this publication stems Professors Palyi, Caglioti and Zucchi have brought together an unusually wide ranging set of opinions about the definition and origin of life. They extend from the creation myths intuited by primitive man to the deductions drawn by modern science. Both attempt in different ways to extract meaning from observation of nature and to predict and control it. The experimental and theoretical tools needed for a deeper understanding of related problems have only been developed over the last few centuries with explosive progress in the last few decades. In contrast, the experimentally unaided human mind has for millennia been grappling with these questions, embodying and institutionalizing speculations in a variety of canonical writs. By virtue of their nature and methodology these must on many points diverge from the views of modern science.

When it comes to the question of definition of life much of the divergence between scientific and spiritual views can be stripped away if one abstains from bringing in anthropocentric arguments. The objections leveled against what is often called the "reductionist" approach is that it leaves out or demystifies the exceedingly complex manifestations of human neuronal activity, thought to require a non-material explanation. If one therefore, as occasionally done at this conference, takes into account that Earth's biomass consists mainly of microbes and addresses the definition of life from their point of view, at least initially and for the sake of clear arguments, the problem of definition and philosophical treatment is greatly simplified. The reason is that this dominant body of life was unknown until the rise of modern science and its manifestations have not been incorporated in religious doctrines or in philosophical concepts such as consciousness or free will. Different ideologies can under these conditions approach the question on an equal basis, drawing on the same material for their resulting revelations and conclusions.

REQUIREMENTS
Much experimental and theoretical work has been published, claiming to model prebiotic chemical processes that would lead to the emergence of life on Earth. The relevance of such claims is limited by cosmochemical and geochemical probability. The conditions to be met may be summarized as:

• sufficient rate of supply of source molecules, competing with their rate of destruction,
• selective concentration of desired reactants at the exclusion of interfering species,
• selective condensation, utilizing natural mechanisms to form sequence specific libraries of nucleotides, peptides or other macromolecules assumed to have formed prebiotic scaffoldings .
• random interreaction and growth of such species to reach the large sizes that eventually permit biofunctionality and beginning natural selection based on fitness. When this point is reached the system emerges from chaos into a darwinian selection regime, governed by rules other than those of chance.

STAGES IN BIOPOESIS
The processes currently speculated to have led to the emergence of life appear in two initial stages. The earliest of these comprises the introduction or generation on Earth (or wherever the development took place) of simple source molecules, from which the construction of more complex entities could be thought to have taken place.

Most schemes for such spontaneous synthesis, some of them modeled in the laboratory, are extremely unlikely to have occurred in nature. There are several reasons for this. One is the common reliance on what now appears as unrealistic assumptions about the physics and chemistry of primitive planetary atmospheres. Such considerations stem from the attempts by Oparin (1924) to understand the environmental conditions necessary to permit the formation of life via peptides, thought at the time to be a crucial element in biopoesis.

As such necessary conditions Oparin inferred an early atmosphere consisting of hydrogen, ammonia, methane and water, conducive to facile formation of a wide range of organic molecules. Prominent geochemists such as V.M.Goldschmidt (1952) raised objections against this assumption, pointing out that the primordial atmosphere in the absence of free oxygen would be dominated by nitrogen, carbon dioxide and water. Volcanic gases have been suggested as another potential source of organic molecules. However, pristine volcanic source gases, in the absence of metallic iron and nitrogen sources in the Earth's crust, consist mainly of carbon dioxide, water, sulfur monoxide and -dioxide . Their lack of primary nitrogen, ammonia and hydrogen carries major geochemical implications. Nonetheless, experiments modeled on global or local hydrogen-methane-ammonia dominated atmospheres continue to convey a distorted impression about the ease with which constructive reactions could have occurred early in Earth's history. It is notable, however, that simple aldehydes, important source compounds for the sugar-phosphate backbones of nucleic acids, do not require hydrogen or methane for formation. Attempts were made in the early part of the last century, in the wake of Woehler, to demonstrate the formation of biomolecules from inorganic sources (biomimetics). Such experiments demonstrated among other products the formation of glycolaldehyde in silent electrical (plasma) discharge reactions carried out in carbon dioxide - carbon monoxide - water atmospheres ( Löb, 1906; 1913; Löb and Sato, 1915).

Aside also from other environmental obstacles such as extreme dilution and need for activation, many difficulties arise in the claim of chemical autosynthetic events, that must be imagined to have led to functional biopolymers. These problems have been succinctly analyzed by Joyce and Orgel (1999) who concluded that the "de novo appearance of oligonucleotides on the primitive Earth would have been a near- miracle." However, in the scientific view the near-miraculous character of the problem is not ascribed to a missing principle, but to our present state of ignorance and limited experimental experience.

Also in the geochemical argumentation there are sufficient uncertainties to admit the finite possibilities of even highly unlikely events or situations in the Archean prebiotic world. We have no preserved record confirming the postulated composition of the atmosphere and hydrosphere during the first 500 million years; such fundamental planetary information may, however, eventually be gleaned from martian sediments . What we know about planetary atmospheric physics and chemistry strongly suggests an atmosphere dominated by nitrogen, carbon dioxide and water, with only minor mixing ratios of reducing molecules. However, we have no proof, a fact utilized by reducing protagonists. Harold Urey, when confronted with obvious geochemical conflicts, used to say "nobody was there to see".

INFORMATION AND MEANING
As pointed out by Yockey (1992;2000) actual messages such as a sequence of synthetic nucleotides or an array of ions in a crystal, may or may not have any meaning (specificity, function). The most fundamental specificity in molecular evolution is catalytic-, particularly autocatalytic function .

For catalytic ligase or replicase activity to appear in sequences of amino acids or nucleotides a relatively large number of residues would need to combine into specific, ordered sequences. The probability for such combinations is small. If such a sequence happens to appear by random combinations in reasonable time (and by natural ligation processes) it will, since lifetime is short, as a rule be an ephemeral event, practically useless unless it is autocatalytic. In p-RNA (Eschenmoser, 1999), autocatalytic size for elongation and replication may be reached already at the tetrameric size, and the production of oligomers is under favorable conditions spontaneous as a consequence of the self- elongating capability of such tetramers . Countervailing effects would arise from excessive dilution, unwanted side reactions with more abundant molecules in the environment, and hydrolysis of the desired products .

Elongation is in itself message- (sequence-) independent; it can proceed with meaningless sequences and produce libraries that consist of mostly meaningless oligomers. The probability of stochastic formation of specific and thus potentially meaningful sequences decreases with increasing length. The size of a ribozyme with demonstrated ligase activity has been shown to be about 100 nucleotide residues (Ekland et al. 1995). If uniqueness of the nucleotide sequence were strictly required the probability for such an oligomer to arise stochastically is approximately 10 exp-158, a highly improbable event. If an additional requirement of regiospecificity of the phospodiester bond would be imposed (Joyce and Orgel, 1999), the probability is further decreased. Reaction of an RNA nucleotide or oligonucleotide with an activated nucleotide normally yields 5'5'-pyrophosphate-, 2',5'-phosphodiester- and 3',5',-linked adducts (Joyce and Orgel, 1999). Without a remedy for this triple complexation, the probability exponent above deteriorates to -614.

Impossibly large improbabilities of this kind loom as a universal threat in the chaos era of chemical evolution. Partial or complete relief from the phosphodiester bond multiplicity is brought about by special reaction conditions (Mueller et al.,1990), favorable catalytic substrate effects such as specific metal ions (Sawai et al. 1988) or mineral surfaces (Pitsch et al. 1989; Ferris and Ertem, 1987) Production of pentose bisphosphate, exclusively with 2,4-phosphodiester bond, was found by Mueller et al.(l.c.) at condensation of formaldehyde and glycolaldehyde phosphate particularly in alkaline solution, and by Krishnamurthy et al, (1999) at condensation of glycoladehyde- and glyceraldehyde phosphate catalyzed by mixed valence double layer hydroxide minerals at neutral pH.

Solace from the tyranny of nucleotide combinatorials is sought in the feeling that strict sequence specificity may not be required through all domains of a functional oligomer, thus making a larger number of library items eligible for participation in the construction of the ultimate functional entity. It is also conceivable that a number of different constructs could have arrived at the same function (G. Joyce, pers.comm.). If the inevitable errors in the vast majority of growing oligonucleotides lead to structures particularly susceptible to inhibition or destruction, this would decrease the risk for exhaustion of the pool of source monomers, but the improbability remains the same for formation of the perfect ultimate functional oligomer. Conversely, if one of the randomly forming molecules would be autocatalytic, chances that it would be contributing toward an ultimate biofunctional species would remain small.

THE FOUR POWERS OF CRYSTALS
Any of the molecules needed for the emergence of a pre-RNA (or pre-Peptide) World would be expected to appear in extreme dilution in the hydrosphere, regardless of extraterrestrial, atmospheric or hydrothermal origin, belying the oft cited concept of "a primordial soup"; a rich broth of all kinds of desirable (and undesirable) organic molecules. In a more realistically dilute state and in competition with hydrolysis, the critical source molecules and intermediate products would be incapable of constructive reactions and would meet death before life. Concentration from a dilute state thus seems to be a prerequisite for the emergence of life. Several geochemical processes have been proposed for this function - eutectic freezing of an aqueous medium (Sanchez et al. 1966), separation into a monomolecular surface oil film, precipitation of solid phases such as ferroferricyanide, Prussian Blue (Arrhenius et al.1993; Keefe and Miller, 1996) and insoluble lead complexes of selected aldoses (Zubay, 1998). Sorption on surface active minerals (Goldschmidt, 1947 [1952]; Bernal, 1947; Cairns-Smith, 1982; Kuma et al.1989) appears as a particularly effective and selective concentration process. Evaporation would be inefficient because of the large initial dilution expected and the destructive effects of the energy source, solar radiation or thermal vaporization.

If a gustatory metaphor were needed to express the disappointing qualities of the 'primordial soup' it would be 'dilute dishwater with a greasy surface film and a bottom deposit of contaminated mud' or 'primordial dishwater' for short, emulating an expression by Lavoisier.

Among the plausible concentration mechanisms, phase separation and surface sorption seem particularly appealing since they also can perform a second necessary function; molecular selection (Pitsch et al.1995, Zubay 1998) Without selection of specific compounds such as aldehydes, phosphates, cyanide, ammonium ion and amino acids, processes directed toward biomolecular precursors would meet formidable obstacles in the form of interference with more abundant reactive products of primordial organic geochemistry such as hydrocarbon derivatives and carboxylic acids. There is also the peril of deleterious cross reaction and thus elimination of desirable ingredients such as by cyanohydrin formation from cyanide and aldehydes (Arrhenius et al., 1997;1998). A number of surface active solid species have demonstrated unique abilities for selective concentration, a property effectively utilized also in chemical technology.

Concentration and selection thus appear as two powerful properties of surface active minerals. The third power resides in their proven ability to catalyze specific reactions between molecules sorbed on external or internal surfaces. Among reactions studied, two appear to be of etiological relevance. One is the selective formation from formaldehyde and glycolaldehyde phosphate of ribose-2,4-bisphosphate, the sugar phosphate backbone molecule in p-RNA referred to above. The reaction proceeds toward completion in a week in 2N NaOH, in two weeks at pH 10.7 and at barely perceptible rate at neutral pH (Mueller et al.,1990). When glycolaldehyde- and glyceraldehyde phosphate are sorbed into the interlayer of the expanding mixed valence Mn-Al hydroxide mineral mangalite from micromolar solution at neutral pH, ribose -2,4-bisphosphate forms as the dominant pentose-2,4-bisphosphate (Krishnamurthy et al. 1999). Tetrose-, mainly erythrose phosphate forms under similar conditions upon dimerization of glycolaldehyde phosphate by mineral interlayer catalysis (Pitsch et al., 1989). Phosphorylation with cyclic triphosphate (trimetaphosphate; Etaix and Orgel, 1978) is also efficiently catalyzed by this heterogeneous mechanism (Kolb et al. 1997).

A potentially significant mineral catalyzed reaction, demonstrated by Ferris and Ertem (1993) applies to a stage further ahead in chemical evolution. It involves the ligation of already formed, activated nucleotides on the crystal edges of the expanding sheet structure silicate montmorillonite. Elongation to 40-mers was achieved with this technique by successive additions of activated nucleotides to the reaction system and there is reason to believe that it, with patience, could be carried to greater length.

A fourth, potentially important property of minerals lies in their aperiodic structure, providing in each individual crystal a specific message - again meaningless in terms of complex biological function but possibly effective in mass production of sequence specific oligomers. It was proposed by Cairns-Smith (1982) that organic oligomeric structures, growing from monomers on mineral surfaces could conceivably be organizationally controlled by aperiodic structural features in the substrate crystal. Cairns-Smith was in this context thinking of the irregular pattern of crystal defects, copied from one layer to the next during crystal growth.

MESSAGES
Ideal, strictly periodic crystals occur only in high school textbooks - all real crystals deviate significantly from simplicity and perfection. Aside from propagated defects, variable ion substitution provides a second type of chemically active sites, distributed in aperiodic patterns. In contrast to defects, these sites strictly obey the geometry of the crystal lattice. With half a dozen or so common substituent ions in surface active minerals such as double layer hydroxides and clay minerals, a variety of patterns are written as meaningless messages on their cleavage planes.

A third type of message is generated by the mixed layer structures characteristic of sheet structure minerals, in which sheets of different composition and therefore with different net charge and thickness may alternate. Typically sheet types alternate in different permutations, A, BB, AA, B, AB, etc. These messages are exposed to the outside world on crystal edges where e.g. the Ferris elongation reaction, referred to above, appears to take place (Franchi et al, 2000).

It should be emphasized that the proposed translation of crystal messages to condensing assemblages of sorbed organic molecules is at this time purely ' the prebiotic chemist's dream' (Joyce and Orgel, 1999). Although it should be experimentally verifiable, this has not yet been achieved. Surface processes, however, enable oligomerization reactions that are necessary for creating the material tape for imprinting of base information. It is then tempting to explore the possibility that the surface distribution of catalytic and sorptive sites may affect the arrangement of the sorbed and catalytically linking monomers. The fact that both crystal sheet surfaces and edges offer two-dimensional sorption lattices does not preclude the predetermined one-dimensional growth of linear oligomers. The surface lattice provides several lineation directions with different spacing between potentially controlling metal cations

Sowerby and collaborators (1998, 1999, 2001) have demonstrated a surface interaction process involving crystal induced ordering. They showed that nucleobases adsorbed on crystal surfaces, e.g. (001) of graphite arrange themselves in a regular pattern and bind preferentially to specific amino acids.

What importance could a crystal-to-oligomer information transfer have had for early molecular assemblages? Innocent of consciousness, free will, and the brutal urges of darwinian selection they must have been struggling aimlessly toward uncharted configurations, a vanishingly small fraction in the end turning out to be endowed with biologically meaningful function. In the presence of aperiodic surface active mineral structures, each microscopic crystallite could be envisaged to produce a large number of oligomers with specific sequences, thus bypassing the low probability of any such sequence being produced by random monomer interaction.

Another contribution by active mineral surfaces is the relief from ambiguity of phosphodiester bonding in an RNA World. The probability of a sequence-specific tetramer being produced randomly and with a prescribed three-fold regiospecificity from four different monomers would be of the order of 12 ! or 10exp-9. If the regiospecificity requirement is relieved by catalytic mineral surfaces as demonstrated for the 2,4-bond in pyranosyl pentose phosphates, the probability is drastically improved to 1 in 24. If furthermore such a sequence could be produced by an originating crystal surface pattern the probability would approach unity. The selective and catalytic power of surface active, ordered solids provided by the mineral kingdom may therefore offer some hope for guidance through molecular chaos toward the state where eventually functionally meaningful messages could appear. If that point were reached, the progressive darwinian laws of selection (Eigen et al. 1971) would take over from the restrictive stochastic laws of chaos.

POWER FAILURES
While the powers 1, 2 and 3 above are well established under idealized conditions in the laboratory, they all leave something to be desired from a geobiological point of view. The fourth power concerns a message, which although it is conceptually tempting, remains scientifically meaningless until it can be verified experimentally. Meanwhile the message has a meaningful and exhortative telomer; "explore".

Arrhenius, G., K.K. Baldridge, S. Richards- Gross, J. Siegel, Glycolonitrile oligomerization: structure of isolated oxazolines, potential heterocycles on early Earth J. Org. Chem. 62, 5522-5525 (1997).

Arrhenius, T., G. Arrhenius, W.J. Paplawsky, Archean geochemistry of formaldehyde and cyanide and the oligomerization of cyanohydrin Orig. Life Evol. Biosphere 24, 1-19 (1994).

Bernal, J.D., The physical basis of life, Proc. Phys. Soc. London, 62, 537-558 (1947)

Cairns-Smith, A.G., Genetic Takeover and the Origin of Life, 477pp (Cambridge Univ. Press, 1982).

Eigen, M., Self-organization of matter and the evolution of biological macromolecules. Naturwissenschaften 58, 465 (1971)

Ekland, E.H., Szostak, J.W. and Bartel, D.P. Structurally complex and highly active RNA ligases derived from random RNA sequences. Science 269, 364-370, (1995).

Eschenmoser, A., Chemical etiology of nucleic acid structure Science 284, 2118-2124, (1999).

Etaix, E., L.E. Orgel, Phosphorylation of nucleosides in aqueous solution using trimetaphosphate: formation of nucleoside triphosphates J. Carbohydrates-Nucleosides-Nucleotides 5, 91-110, (1978).

Ferris, J.P. and G.Ertem, Montmorillonite catalysis of RNA oligomer formation in aqueous solution. A model for prebiotic formation of RNA. J.Am.Chem.Soc. 115, 12270- 75 (1993)

Ferris J.P., Hill A.R., Liu R., and Orgel L.E. Synthesis of long prebiotic oligomers on mineral surfaces. Nature 381, 59-61, (1996).

Franchi, M., E.Bramanti, L. M. Bonzi, P.L. Orioli, C.Vettori and E. Gallori, Clay-nucleic acid complexes: characteristics and implications for the preservation of genetic material in primeval habitats, Orig. Life Evol. Biosphere 29, 297-315, (1999).

Goldschmidt, V.M., Geochemical aspects of the origin of complex organic molecules on the Earth, as precursors to organic life (1947). Posthumously published by N.W. Pirie, ed; New Biology 12, 97-105, (1952)

Joyce, G.F., L.E. Orgel, Prospects for understanding the origin of the RNA World. In The RNA World II, T. Gesteland et al., eds.(Cold Spring Harbor Laboratory Press, 49-77, 1999).

Keefe, A.D., S.L. Miller, Was ferrocyanide a prebiotic reagent? Orig. Life Evol. Biosphere 26, 111-129 (1996)

Kolb, V., S. Zhang, Y. Xu, G. Arrhenius, Mineral-induced phosphorylation of glycolate ion - a metaphor in chemical evolution, Orig. Life Evol. Biosphere 27, 485-503 (1997).

Krishnamurthy, R., S. Pitsch, G. Arrhenius, Mineral induced formation of pentose-2,4-bisphosphates, Orig. Life. Evol. Biosphere 29, 139-152 (1999).

Kuma, K., B.Gedulin, W.Paplawsky, G. Arrhenius, Mixed-valence hydroxides as bioorganic host minerals, Orig. Life Evol. Biosphere 19, 139-152 (1989).

Löb, W., Studien über die chemische Wirkung der stillen elektrischen Entladung Zeitschrift für Elektrochemie 11, 282-316 (1906).

Löb, W.,Über das Verhalten des Formamids unter der Wirkung der stillen Entladung. Ein Beitrag zur Frage der Stickstoff-Assimilation Ber. d. Deutschen Chem. Ges. 46, 684-697 (1914).

Löb,W., A.Sato, Zur Frage der Elektrokultur, Bioochem.Ztschr. 69, 1-34 (1915)

Mueller,D., S.Pitsch, A.Kittaka, E.Wagner, C.E.Wintner, A.Eschenmoser, Chemie von alpha-Aminonitrilen; Aldomerisierung von Glycolaldehyd-phosphat, Helv. Chim. Acta 73, 1410 (1990)

Oparin, O., A.I., Proiskhozdenie Zhizny (Moskovskii Rabochii, Moscow, 1924).

Pitsch, S., A. Eschenmoser, B. Gedulin, S. Hui, G. Arrhenius, Mineral induced formation of sugar phosphates Orig. Life Evol. Biosphere 25, 294-334 (1995).

Sanchez, R.A., J.P. Ferris, L.E. Orgel, Conditions for purine synthesis: Did prebiotic synthesis occur at low temperatures? Science 153, 72-73 (1966).

Schöning, K-U., P.Scholz, S.Guntha, X.Wu, R.Krishnamurthy, A.Eschenmoser., Chemical etiology of nucleic acid structure: The alpha- threofuranosyl-(3'->2')-oligonucleotide system Science 290, 1347-51 (2000).

Sowerby,S.J., W.M.Heckl, The role of self-assembled monolayers of the purine and pyrimidine bases in the emergence of life. Orig.LIfe Evol.Biosphere, 28, 283-310 (1998)

Sowerby, S.J., G.B.Petersen, Scanning tunneling microscopy and molecular modeling of xanthine monolayers, self-assembled at the solid-liquid interface: Relevance to the origin of life. Orig.Life Evol.Biosphere, 29, 597-614 (1999)

Sowerby, S.J., G.B. Petersen and N.G. Holm (2001). Primordial Coding of Amino Acids Submitted

Sowerby, S.J., C.A. Cohn, W.M. Heckl and N.G. Holm (2001). Differential adsorption of nucleic acid bases: Relevance to the origin of life, Proc.Natnl Acad.Sci.USA, 98, 820-822 (2001)

Yockey, H.P.,Information Theory and Molecular Biology (Cambridge University Press, Cambridge, UK,

Yockey, H.P., Origin of life on earth and Shannon's theory of communication, Computers and Chemistry, 24, 105-123 (2000).

Zubay, G. Studies on the lead-catalyzed synthesis of aldopentoses, Orig.LIfe Evol.Biosphere, 28, 13-26 (1998)

©Copyright 2002 Arrhenius Laboratory - All Rights Reserved