Darwin’s warm little pond revisited: from molecules to the origin of life
- First Online:
- Cite this article as:
- Follmann, H. & Brownson, C. Naturwissenschaften (2009) 96: 1265. doi:10.1007/s00114-009-0602-1
All known cosmic and geological conditions and laws of chemistry and thermodynamics allow that complex organic matter could have formed spontaneously on pristine planet Earth about 4,000 mya. Simple gasses and minerals on the surface and in oceans of the early Earth reacted and were eventually organized in supramolecular aggregates and enveloped cells that evolved into primitive forms of life. Chemical evolution, which preceded all species of extant organisms, is a fact. In this review, we have concentrated on experimental and theoretical research published over the last two decades, which has added a wealth of new details and helped to close gaps in our previous understanding of this multifaceted field. Recent exciting progress in the molecular and genetic analyses of existing life, in particular microorganisms of ancient origin, even supports the possibility that a cellular, self-reproducing common ancestor might be assembled and resurrected in anaerobic cultures at some time in the future. Charles Darwin did not, and indeed, could not, address and specify the earliest phases of life which preceded the Origin of Species. However, in a famous letter, he sketched “a warm little pond with all sorts of… (chemicals, in which) …a protein was chemically formed.” We try to trace the impact of his charming clear-sighted metaphor up to the present time.
KeywordsCharles Darwin Chemical evolution DNA and RNA Iron–sulphur world Last universal ancestor Organic soups Origin of life Thermodynamics
It is often said that all the conditions for the first production of a living organism are now present which could ever have been present. But If (and oh what a big if) we could conceive in some warm little pond with all sorts of ammonia and phosphoric salts, light, heat, electricity etc. present, that a protein compound was chemically formed, ready to undergo still more complex changes at the present such matter would be instantly devoured, which would not have been the case before living creatures were formed.
Charles Darwin (1809–1882) wrote these sentences in a letter to his botanist friend Joseph D. Hooker in 1871. What an amazing insight, decades before phosphorus was recognized as an essential element of cellular energy metabolism and heredity, although it had been discovered at the same time in “nuclein” (i.e., DNA; Miescher 1871). It was also long before amino acids, the nitrogen-containing monomers of proteins, were formed in Stanley Miller’s seminal “organic soup” experiments conducted under the influence of heat, electricity, and light (Miller 1953, 1955).
Charles Darwin was an outstanding naturalist, versed in biology, husbandry, geology, and mineralogy. Was he also a secret, visionary chemist? We do not really know. Charles and his brother Erasmus studied the “Chemical catechism” of Samuel Parkes and performed chemical experiments in a little laboratory in their father’s garden during their younger years. The formation of gasses from minerals (for example, carbon dioxide from marble and acids, or ammonia liberated from ammonium chloride with CaO or quick lime) interested Charles greatly, to the extent that at school he was nicknamed “Gas.” During his medical studies at Edinburgh University (1825–1827), at the age of 17, Charles enjoyed the modern chemistry lectures based on Dalton’s theory of atoms taught by Thomas C. Hope. Indeed, Darwin remembered these as the only interesting lectures at Edinburgh. He did not specifically mention chemistry during his years at Cambridge (1828–1831). Later in his life, Darwin exchanged letters with almost 2,000 people and many of them (contained in the Darwin project database; Darwin Correspondence Project, http://www.darwinproject.ac.uk) addressed properties and reactions of lime, carbonates, ammonia, nitrogen in bones, etc. He knew about Justus Liebig, the famous German chemist with a reputation for analyses of the chemical composition of soils, food, and other biological matter which flourished in the nineteenth century. Thus, it would appear that Charles Darwin had acquired a good knowledge of contemporary chemistry.
One can only speculate whether twentieth century chemists were inspired by Darwin’s Warm Pond scenario when they set out to discuss and explore the chemical evolution of life in laboratory flasks (Oparin 1936; Urey 1952; Miller 1953, 1955). The above letter to Hooker is quoted by Oparin, Hardin (1950), and Calvin (1969); but as far as we could ascertain, it was not mentioned in the early experimental contributions from other authors. It is a bit curious that Darwin did not mention a carbon source (non-negotiable for organic compounds), but he was probably not aware of the biological significance of methane (marsh gas) in anaerobic places. On the other hand, our colleagues of 50 years ago did not comment on phosphorus, equally important for life. Oh what inciting questions remain for historians!
In his famous Origin of Species (1859), Charles Darwin does not reflect upon the earliest origins of life, except for one brief sentence in chapter XIV. In a late letter (1881) to the botanist Nathaniel Wallich, he wrote “You expressed quite correctly … that I had intentionally left the question of the Origin of Life uncanvassed as being altogether ultra vires in the present state of our knowledge.”
About a century of research later, Sol Spiegelman (Mills et al. 1967) and Manfred Eigen (1981) demonstrated both experimentally and theoretically that the Darwinian principles of replication, variation, and selection constitute a law of nature applicable to populations of molecules such as nucleic acids during self-reproduction, as well as to organisms. With the present knowledge of chemistry and molecular biology, the Darwin Jubilee Anniversary 2009 is worthy of an in-depth review of our molecular precursors in an enveloped, cell-like “species” ready to proceed on the long march towards more complex forms of life.
What is “chemical evolution”? It describes the chemical processes that took place on the prebiotic Earth about 4,500 to 3,500 mya. These events preceded biological evolution, a phase which led to the appearance of first living cells capable of self-reproduction at the expense of some rudimentary metabolism. Scientific concepts and laboratory experiments aimed to link life and its material constituents to the laws of chemistry and physics date back almost 100 years (Löb 1913; Oparin 1924). They gained momentum after Stanley Miller reported his now-classic results in 1953 (made public in Germany in the journal Angewandte Chemie in the same year) and thereafter. During more recent years, inspiring new ideas and experiments opened new “worlds” such as the iron–sulphur world of prebiotic chemistry and the RNA world of early genetics and catalysis. Today, molecular genetics and bioinformatics contribute deep-rooted sequence trees of all living organisms from which the characteristics of a potential last universal cellular ancestor (LUCA) have emerged.
Individuals who take an interest in evolution and the origins of life hope and, indeed, expect that modern science will provide a detailed and conclusive description. In a scenario of such complexity, when dealing with one’s own descent, gaps and uncertainties are uncomfortable and unwelcome. Those of us in science who have touched on this multifaceted matter in theory or practice know that this expectation cannot be met despite the great progress in our knowledge made in recent decades. Authors of review articles and textbooks are not always immune to streamlining the arduous pathways from molecules to cells and organisms. In order to better explain the underlying principles of these pathways, specialists in cell physiology, chemistry, enzymology, or genetics may be tempted to belittle or disregard the problems with certain facts in scientific fields other than their own. This can be a disadvantage when evolution in general is questioned by religious dogmatists.
In this review, we have attempted to avoid bias and to describe diverse or conflicting results where necessary. It is not possible, nor intended, to trace our current understanding of the genesis of life on Earth in full. There have been many hundreds, if not thousands, of chemical and physicochemical experiments and theoretical analyses which have been performed under “warm little pond conditions” and in other environments. Some of the early work remained incomplete or inconclusive because of analytical and other experimental limitations but can now be assessed with modern methodology. We shall point to crossroads where proponents and opponents of a specific pathway (reducing or redox-neutral conditions? solution or surface chemistry? protein or RNA world?) have difficulties in defining an outcome.
Information presented here places emphasis on studies published during the past two decades and only gives the essence of earlier work. Numerous books and reviews (e.g., Urey 1952; Oparin 1957; Calvin 1969; Lemmon 1970; Eigen 1971; Kvenvolden 1974; Follmann 1981; Kutschera and Niklas 2004; Kutschera 2009) describe in more detail previous experimental studies and theories of chemical and biological evolution.
Chemistry on a pristine planet Earth
Life, as viewed by chemists and physicists, describes highly organized entities (“cells”), composed of various kinds of organic and inorganic molecules and macromolecules, enclosed and separated from the environment by hydrophobic membranes. Moreover, the definition of life requires cells to acquire both energy (i.e., light, or energy-rich molecules) and new organic matter (i.e., carbon compounds and other nutrients) from outside in order to build and feed a genetic apparatus capable of reproducing the system. Finally, transmissible variations (mutations) of the genetic systems are necessary for the continued evolution of vital activities. All these properties demand very special environmental conditions.
A habitable place
Did “bio”chemistry start on Earth or was life brought in from space? Earth-bound chemical reactions and products, such as metabolites, life-like polymers, and cellular precursors, will be at the centre of this review. Nevertheless, large radio telescopes have detected many simple organic molecules in cosmic clouds, and analyses of carbon-containing meteorites that have fallen to the Earth have identified many organic compounds, in particular amino acids. It is conceivable that the European Rosetta space mission and its “cometary sampling and composition” COSAC instrument may find chirality—a characteristic of biomolecules—on arrival at the small comet Churyumov-Gerasimenko, 450 million kilometers away, in the year 2014 (Goesmann et al. 2007). Such findings from space deserve attention and are described below where pertinent but they do not prove that life originated “out there.”
Earth is the only planet within the solar system on which the above-mentioned physical conditions for life exist. It is unique because its mean surface temperature, at a distance of 150 × 106 km from the glaring Sun, allows water to exist in all three physical states (as solid ice, liquid, and vapor). This is not the case on our smaller neighbor planets, hot Venus and cold Mars. Liquid water is a superior polar medium that dissolves, transports, and allows reactions between organic and inorganic compounds inside and outside cells. In a gaseous atmosphere, water vapor and gaseous carbon, nitrogen, and sulphur species and their various products coexist in numerous equilibria. Ice can trap, concentrate, and conserve complex products of primordial syntheses. Additionally water and abiotic hydrophobic lipids constitute biphasic media. When agitated by wind or waves, partition equilibria and compartmentalization processes in these media can separate different classes of newly formed and future biomolecules of different polarities.
Astronomers have, in recent years, detected many “exoplanets” and planet-forming clouds of gasses and dust which circle around stars in faraway galaxies (e.g., Beaulieu and PLANET Collaboration 2006; Eisner 2007). Most such planets are of giant Jupiter size but one with only 5.5 Earth masses size was also described. It is unanswerable at this time whether any of these bodies could support life elsewhere in the universe, but future infrared-based remote sensing technologies and a search for key molecules, including water, may someday provide an answer (Bada 2001). More data are expected from the National Aeronautics and Space Administration’s current Kepler spacecraft mission.
Analyses of numerous long-lived radioactive and stable isotopes (of hafnium, tungsten, uranium, thorium, lead, zirconium, and others) in meteorites and terrestrial rocks indicate that the Sun and its planets accreted from gasses and cosmic debris 4,600 mya (Halliday 2001; Jacobsen 2003). It is assumed that the primitive solar nebula contained matter of a much older star which had collapsed and exploded as a “supernova.” This event must have been significant for life as some contemporary proteins require catalytic metals heavier than iron (e.g., zinc or molybdenum) for catalytic activity. These were not produced by thermonuclear reactions in the young Sun but they came with the remains of the supernova. The fact that atoms such as essential molybdenum, found in our digestive enzyme xanthine oxidase, are direct descendants of the very origin of the universe is an intriguing consequence of this event.
The early Earth must have been a violent and chaotic body of rocks and gasses, with an initial surface temperature of around 1,000°C, and covered with huge impact craters from collisions with large and small asteroids during the “heavy bombardment” phase. Most of the lightest gasses, hydrogen and helium, escaped rapidly, and it is believed that the planet had solid rocks and an atmosphere (see below) between 4,400 and 4,100 mya. When temperatures dropped below its critical point (374°C), water condensed and filled crater lakes and early oceans; this occurred an estimated 4,000 million years before present. Indeed, one of the oldest sedimentary rocks (the Isua cherts of West Greenland) is 3,800 million years old.
Carbon of biogenic origin and presumed cellular microfossils has been reported in these and other 3,500-million-year-old sediments (Pflug and Jaeschke-Boyer 1979; Schidlowski 1988; Schopf 1993; Hayes 1996). While these objects were analyzed by powerful microscopic and microchemical techniques, doubts have remained and artifacts cannot be excluded (McLoughlin et al. 2007). Nevertheless, it appears safe to say that “Darwin’s dilemma,” i.e., the apparent missing Precambrian fossil record is no longer a problem (Schopf 1999).
It is generally agreed that the evolution of life—aggregation of fragile organic molecules—required a liquid, aqueous medium. (Thales of Miletus told us that at about 600 B.C.!) At the outset, however, “ecosystems” of accumulated organic matter would have been annihilated on a global scale in one of the frequent asteroid bombardments that continued to about 3,900 mya (Sleep et al. 1989). For the establishment of life, the time between such events had to be greater than the time required to evolve living organisms. It has been estimated from analyses of impact craters and energies (Oberbeck and Fogleman 1989) that such conditions would have prevailed between 4,000 and 3,700 mya. The agreement between these dates and those quoted above for the presumed existence of very ancient microfossils is remarkable. New numerical models of the habitability of the early Earth have just been published (Abramov and Mojzsis 2009).
Living matter is typically composed of 15–20 chemical elements, including six indispensable macroelements (C, H, O, N, P and S) plus the more or less essential microelements (Na, Mg, Cl, K, Ca, Mn, Fe, Co, Ni, Cu, Zn, Mo, and W). Although some of these elements are rare or sparingly soluble, their availability and uptake from the atmosphere, water, and minerals has obviously not limited chemical and biotic evolution in general. Elements capable of existing in different valence states (C, N, O, S, iron, and other transition metals) will have predominated on the pristine hydrogen-rich planet Earth in low, reduced valences: C as methane, N as ammonia or nitrogen, O in water, S as sulfide but not sulfate, and iron in bivalent oxides or sulfides (FeO, FeS) or even as a metal. The global redox potential began to change when some oxygen was liberated by UV light-induced radiolysis of water and much more drastically after the appearance of oxygenic photosynthesis in primitive bacteria (the later endosymbionts of green algae). The period during which the Earth’s atmosphere transmuted from a reduced to a neutral and an oxidizing state is conspicuous in geology for the precipitation of massive red iron oxides (“banded iron formations”) up to 2,000 mya. In biology, it opened the transit from nitric oxide (NO), nitrate, and sulfate as electron sinks (Ducluzeau et al. 2009) to modern oxygen respiration, thereby supplying ever more metabolic energy for hungry Proterozoic cells. We will not elaborate these dramatic biochemical changes further.
It is not widely known that Charles Darwin’s “warm little pond” metaphor sparked a stimulating study (Spaargaren 1985) on the origin of the inorganic elements of life: From oceans? rivers? or panspermia from space? The latter author noted that the elemental composition of contemporary organisms, marine or terrestrial, is remarkably conservative from bacteria and fungi to fish and man. This is possibly dictated by the chemistry of all universal biochemical reactions taken together and is, therefore, still similar to the composition of early cells. The proportions of macroelements and microelements in organisms differ from those in seawater but the composition of river water is reasonably comparable with living matter—suggesting that organisms of such elemental composition originated in warm, evaporating little ponds fed by rivers. It can be concluded that the elemental profile of earthly life does not support Arrhenius’ hypothesis of seeds of life propelled through space by light.
Energies at the surface and in the upper layers of early Earth, cf. Miller and Urey (1959) for quantitative estimates
Standard energy source for chemistry; allows evaporation and sublimation; destructive if too high
Early large impacts
Substantial in earliest atmosphere
Dominant, fainter in the beginning; day-night cycles
≈ 1 % of total
High energy, photochemistry
Absorbed by pigments, emerging photosynthesis
Stronger than present
Allows surface reactions and dry-wet cycles
Hot lava, hot springs
Local (uranium, thorium, K-40)
Uncertain, maybe in singular events; may induce chiral discrimination
Ubiquitous, stronger than present
Excitation of molecules, generation of reactive intermediates (ions, radicals)
High energy; contributions uncertain
Common in chemistry
Driving other reactions, e.g., in iron–sulphur chemistry, redox reactions, etc.
Again, this gives us every reason to muse about Charles Darwin. Of the seven energy entries in Table 1, he could not have had any knowledge of natural radioactivity or cosmic radiation as these were discovered much later. He would also have been unaware of the abstract laws of thermodynamics (elaborated in depth by J.W. Gibbs only in 1875). Nevertheless, he does identify all the energies that would potentially drive evolution in a warm little pond, viz. heat (including volcanic environments, not specifically mentioned), light, and electricity!
Towards biogenic molecules
Ammonia, important for abiotic amino acid formation, is considered a component of the earliest atmosphere and/or was dissolved in the oceans. If it was depleted at some time or in a particular location, the catalytic photoreduction of nitrogen gas by UV light-driven water dissociation, in the presence of titanium dioxide or iron minerals, was a rich alternative source (Schrauzer et al. 1983; Brandes et al. 1998). Another reactive component, hydrogen cyanide (H–C≡N, prussic acid), is formed from methane and reduced nitrogen species in the gas phase and is abundant in volcanic exhalations. Huge amounts (>10 million tons per year) of it may have been produced during impact–shock events (Fegley et al. 1986; Chyba and Sagan 1992). Below 0°C, HCN and water exist in an HCN-rich eutectic state which appears ideally suited for condensation reactions and freezing-out of heterocyclic products.
“Fluctuating environments” are indeed a recurring theme in biogenesis research both in the laboratory and in theory. These include day and night, summer and winter, faint or bright Sun, hot and cold, wet and dry, and dissolved or mineral-adsorbed. Primitive chemistry on Earth differed in the atmosphere, flooded ocean basins, warm little freshwater ponds, submarine hot springs and on hot volcanic rocks, and in many other locations depending on the source of energy (Table 1), temperature, reaction times, presence of inorganic catalysts, or precipitation of products from solution. It is often wrongly presumed that educt concentrations, temperatures, and other variables of a particular prebiotic synthesis were the same and equally productive worldwide. It is not difficult to imagine that however varied the scenarios and the inventory of chemicals, progress towards ever more complex chemical structures, supramolecular aggregates, and “natural selection” cycles became inevitable during the hundreds millions of years.
It has recently been possible (Gaucher et al. 2008) to derive the trend of temperatures in Precambrian oceans for the time after life was established by measuring and comparing the thermostability of ancient ribosomal proteins. Obviously, it would not be possible for any cell to exist under conditions where its ribosomes, delicate nucleoprotein particles, would denature and disintegrate. From these studies, it was concluded that the aqueous environment supporting ancient life changed from a hot start (≥80°C), with thermophiles or even hyperthermophiles, to lukewarm (30°C), which was suitable for mesophiles in the Precambrian period between 3,500 and 500 mya.
Ponds of organic soup and rocky shores with organic scum
Selected classes of organic, “biogenic” compounds and structures obtained in prebiotic Earth simulation experiments
Type of chemistry
Products, yields, comments
Amino acids and related amino compounds
Electric dischargesa, heat UV light, HCN, CO Also in ice
Abundant, numerous ways of formation, (reducing or neutral); Simulating interstellar space
Alcohols and carboxylic acids
Carbohydrates (sugars) Glycerophosphate
Formose reaction (HCHO) + phosphate, cyanamide
Borate minerals stabilize → phospholipids (membranes)
More detailed studies desirable
Peptide bond formation
Heat, clay, salt-induced, Iron-sulphur, carbonyl sulfide
Many condensation reactions
Apatite, carbonyl sulfide, heat
More studies desired for more efficient conditions
Accompany amino acids and HCN
Low yield, characterized by spectra and metal complexes
Fischer–Tropsch + NH3, CO–N2
Aldehydes + amino-oxazole
Cytidine and uridine phosphates
Purine and pyrimidinePolyribonucleotides
Evaporation with Mg, Ca, Zn salts
Best yield in presence of complementary strands
More study necessary
Sulphur compounds, thiols
Iron–sulphur; H2S addition to olefins
Various metabolites, activated intermediates
Iron–sulphurb and other chemical environments, often in low yield
Increasing numbers described
Amino acids, together with amines, are the most frequent products formed in reaction mixtures of reduced carbon and nitrogen species as in Miller’s first prebiotic experiments and in the variants used by other workers (e.g., Bossard et al. 1982; Miyakawa et al. 2002). These observations are in obvious accord with the ease and large yields of classical syntheses of α-amino acids in the laboratory and with the stability of amino acids due to their zwitterionic structure. Amino acids are also found in the organic fraction extractable from carbonaceous meteorites (see below). The number of amino acids which have been characterized over the years is large. Small fractions of additional isomers have been identified in the preserved, dried residues of Miller’s early experiments (Johnson et al. 2008), bringing the total number of different products up to at least 30. Unexpectedly, in contrast to hot reaction conditions intense UV irradiation of a CO2:CO:CH3OH:H2O:NH3 ice matrix frozen in a vacuum at 12 K (−261°C) produced 20 amino acids and other nitrogenous compounds (Munoz Caro et al. 2002). These experiments served to simulate the interstellar medium for comparison with future results expected from the COSAC cometary mission.
Only recently have the circumstances of amino acid synthesis under redox-neutral conditions been explored, based on the assumption that the atmosphere lost its reducing character around 4,000 mya. It was found that product formation also occurred under electric discharges in a neutral mixture of carbon dioxide, nitrogen, and water vapor above liquid water with or without sodium chloride (Plankensteiner et al. 2006) and that the yield of amino acids was greatly increased in the presence of oxidation inhibitors such as ferrous ions (Cleaves et al. 2008). Thus, neutral, hydrogen-depleted atmospheres may have been more important and productive for prebiotic chemistry on Earth than previously thought.
Amino acids dominating under abiotic, early biotic, and contemporary conditions
In a carbon meteoritea
In sterile volcanic lava
In abiotic experiments
Early triplet code codonsb
In average modern proteins (%)
Valine + leucine
Leucine, valine (7 each)
Phe + tyrosine
Phe + Tyr (7)
Sugars are hydrophilic, conformation-determining essential parts of nucleic acids and many coenzymes and metabolites. (Their important role in extant plant life and energy metabolism is not discussed as it is not relevant here.) Investigations of prebiotic sugar chemistry have had their difficulties. “Useful” C5 and C6 species (such as ribose or glucose) form from C1 and C2 precursors (formaldehyde CH2O or glycolaldehyde) in the base-catalyzed “formose reaction” (Pfeil and Ruckert 1961; Follmann 1986) but poor yields, complex isomeric mixtures, and limited chemical stability have not been encouraging. Furthermore, the availability of sufficient prebiotic formaldehyde had been disputed but more recently a rich source has been identified in large atmospheric impacts (Fegley et al. 1986). Another avenue is aldol condensation of glyceraldehyde (C3) catalyzed by Fe(OH)O that produces fructose and other hexoses (Weber 1992). Finally, a new pathway exploiting the well-known esterification of sugar cis-diol structures with boric acid (borate minerals; Ricardo et al. 2004) now permits pentose formation in good stability and yield.
Hydrocarbons, long-chain alcohols, and carboxylic acids
Organic molecules with long linear or branched hydrophobic carbon chains are indispensable for the integrity of cells and intracellular compartments in aqueous environments. Present-day biosyntheses are complex and of little use for prebiotic simulation. Fischer–Tropsch type reactions using iron catalysts (long-known in technical coal chemistry) can produce numerous compounds from CO/H2/H2O gas mixtures at around 250°C with product spectra that often resemble extracts from Precambrian cherts (cf. Kvenvolden 1974). Hydrophobic surface layers can also be seen when the product mixtures of Miller-type sparking experiments cool down. Hydrocarbons are photochemically produced from liquid methane even in the very cold atmosphere of Titan, the largest moon of Saturn (Porco et al. 2005).
Heterocyclic bases and nucleosides
The state of the art changed very recently when an efficient synthesis of cytidine and uridine phosphates was published which bypasses free ribose and pyrimidine bases as precursors (Powner et al. 2009). In an ingenious combination of plausible prebiotic reactions, the authors obtained pyrimidine ribonucleoside 2′,3′-cyclic phosphates from glycolaldehyde and glyceraldehyde (i.e., sugar components) plus cyanamide and cyanoacetylene (forming 2-amino-oxazole as a heterocyclic key intermediate) plus phosphate ions (which serve as general acid/base catalyst and pH buffer). This reaction sequence and the promotion of abiotic sugar formation by borate minerals mentioned above are outstanding examples of how new ideas and synthetic efforts presently expand and will eventually complete our knowledge of prebiotic chemistry.
Phosphate esters and anhydrides
If one small molecule was to be named a biochemical symbol of present-day life, adenosine triphosphate (ATP) would be the obvious candidate. It links energy generation and storage with biosyntheses and provides the phosphate ester links in nucleic acids and many coenzymes. (Remember that phosphoric salts were destined to spice the warm little pond!) The problem in early days must have been that mineral orthophosphates are almost insoluble and are unreactive in water and that the energy-rich anhydride bonds of diphosphates and polyphosphates which are needed for organic syntheses undergo slow hydrolysis. Nevertheless, calcium and magnesium diphosphates (Ca2P2O7, or Mg2P2O7 formed from MgHPO4 under hydrothermal conditions) are good candidates for phosphorylation (Seel and Schinnerling 1978; Seel et al. 1985). Heating calcium phosphates (apatite) in the presence of acidic basaltic rocks to high temperatures (Yamagata et al. 1991) produces volatile phosphorus oxide (P4O10) which in water yields soluble diphosphates, triphosphates, and polyphosphates capable of phosphorylating alcohols, sugars, etc. The significance of such volcanic heat simulation has been questioned because of low yields and competing hydrolytic reactions (Keefe and Miller 1995); unfortunately, until now, the phosphorylation potential of P4O10 samples generated in this way has not been tested. More recent experiments (Leman et al. 2006) show that inorganic phosphate and amino acids in the presence of carbonyl sulfide (O=C=S), a component of volcanic gas emissions, combine to form aminoacyl phosphates which in turn can serve as phosphorylating agents and promote phosphoryl transfer plus peptide synthesis under mild aqueous conditions.
These results may mark a breakthrough in prebiotic phosphate utilization. Nevertheless, a hypothetical, alternative route via more reduced and more water-soluble phosphites (Pasek 2008) should also be experimentally explored.
Chemistry on surfaces
Many of these laboratory reactions remind us of the metabolic traits and metal–sulphur centres known from chemoautotrophic thermophilic bacteria and archaea of ancient origin. Little doubt remains that the FeS/FeS2 system once powered a productive surface metabolism on early Earth where abundant iron sulfide membranes existed near hydrothermal vents (Russell et al. 1994). It is evident that this system contributed to primeval carbon fixation and many different life-associated molecules.
Warm organic soup or a hot mineral paste? It has been argued that the organic broth theory and the iron–sulphur world theory of the chemical origins of life are incompatible (Wächtershäuser 2000). We see no reason for such a rigorous interpretation. Both scenarios are supported by robust experimental data and both could have found ample room in different locations on a highly structured, compartmentalized, unstable surface of the early Earth. Initially, their coexistence could even have been an advantage as newly formed anionic compounds (carboxylic acids, organic phosphates) preferred the positive pyrite surfaces whereas cationic products (amines, heterocyclic bases) accumulated on aluminosilicate sediments with negatively polarized surfaces. Frequent local mixing of two different compartments, e.g., in earthquake-rich regions, is easily imaginable. It should also be noted that the aggregation of “true” cellular structures from lipids, macromolecules, etc. most probably required mesophilic and not (hyper)thermophilic conditions (Miller and Lazcano 1995; Gutfraind and Kempf 2008). Moreover, it is questionable whether iron–sulphur chemistry is able to generate all the different kinds of chemicals required for a protocell (Keefe and Miller 1995).
Besides their catalytic action, another general property of mineral surfaces, their capacity to adsorb and activate different molecules next to each other, lowers activation energies and thereby favors bimolecular condensation or other reactions. Abundant surface-rich clay minerals have, therefore, been studied for a possible role in prebiotic syntheses (Paecht-Horowitz 1976; Paecht-Horowitz and Eirich 1988). It has been demonstrated that kaolinite enhances the thermal condensation of glycine to oligopeptides (Lahav et al. 1978). Montmorillonite, a layered aluminosilicate with excess negative surface charge adsorbs amino acid adenylates (the starting substrates of biological, ribosomal peptide synthesis) and enables production of homomeric or heteromeric peptides of up to 40 amino acids chain length at neutral pH and ambient temperature. Other authors have succeeded in generating long oligonucleotides from an activated adenosine phosphate on mineral surfaces (Ferris and Ertem 1993; Ferris et al. 1996). These reactions with complex kinetics deserve further and more intensive study.
Organic molecules in space
Let us finally consider the often-proclaimed arrival on Earth of more or less developed life from space. (Some people, including scientists, believe that life should not and, therefore, could not have originated on humble little planet Earth.) Panspermia or comet nuclei harboring dormant life spores, including DNA, may be charming scenarios but do not deserve much attention. Such “seeds of life” would never have survived the dehydrating vacuum, low temperature, and intense UV and cosmic radiation on their way to Earth (Weber and Greenberg 1985). Nevertheless, extraterrestrial organic matter has always and still does reach Earth, together with silicates and metals, in interplanetary dust particles (IDPs) and in one category of meteorites, the carbonaceous chondrites (see below). A concise analysis of the fluxes of inorganic carbon, nitrogen, and phosphorus from such sources on the early Earth has just been compiled (Pasek and Lauretta 2008). Estimates are up to five million tons of imported “biogenic” elements per year during the time of heavy bombardment which will have mixed with primordial terrestrial matter. As these events happened before and during the earliest phase of complex organic syntheses on the cooling planet, it makes no sense to differentiate between terrestrial and “outside” precursor molecules. After the heavy bombardment had ceased, the influx of IDPs dropped and today only around 100 tons of C, N, and P per year reach the surface of Earth.
One more extraterrestrial, extragalactic source of chemicals should not be forgotten. A multitude of inorganic and organic molecules in the interstellar clouds of our Milky Way galaxy have been identified by radio and infrared astronomy (Zuckerman 1977; Millar 2004). The surveys of these clouds now include more than 100 species ranging from diatomic and triatomic molecules (hydrogen, water, hydrogen cyanide, nitrous oxide, carbonyl sulfide, formic acid, etc.) up to an exotic 13-atom pentaacetylene nitrile, HC11N. With ethyl alcohol (C2H5OH), it even boasts a molecule of special popularity among us earthlings. It should, however, be noted that their density in space is typically not more than ten atoms per cubic centimeter and that radio telescopes on Earth do not collect and concentrate such matter from a million light-years away but merely record the signatures of rotational and vibrational spectra. There is no opportunity for specific contributions of any such molecule to life.
Natural selection among molecules
One of the most disputed questions concerning the Origin of Life is the way in which chirality arose in the biogenic amino acids which in turn constitute chiral peptides and proteins. Molecules in which a tetrahedral carbon atom carries four different substituents (an asymmetric or chiral atom) exist in pairs of stereoisomers. d-Alanine and l-alanine (Fig. 5) or d-lactic acid and l-lactic acid (with –OH in place of –NH2 groups) are simple examples. The two “enantiomers” are sterically mirror images of one another, but very similar in chemical behavior, with only tiny differences between them (see below). However, random distribution of both in a polymer (e.g., an enzyme) would produce a set of irregular three-dimensional structures that would be unfavorable for specific functions (e.g., substrate specificity). Such disorderly polypeptides are not found in nature where all the amino acids, incorporated into peptides and proteins, are of the same l form. Chirality is also common among sugar molecules, many of which contain several chiral carbons. A hexose sugar with n = 4 asymmetric C can occur in 2n = 16 stereoisomers (and all have been synthesized in the laboratory) but only three of them (d-glucose, d-galactose, and d-mannose) are of biochemical significance. The sugar moiety of nucleotides is always d-ribose or deoxyribose. From all these known coherent properties of life, we can reasonably conclude that natural selection did occur among prebiotic organics from the beginning.
Which subtle chemical and physical forces could establish and augment differences between the stereoisomers of a racemic d,l mixture and produce an enantiomeric excess (ee) of one over the other (Thiemann 1981)? The predominance of chirality or “handedness” of amino acids and other organics remains a difficult problem still today.
The question “Is the Solar System inherently left-handed?” (Chyba 1997) was prompted by the continuing analyses of the big Murchison meteorite, a carbonaceous chondrite which fell in Australia in 1969. It was quickly recovered and shielded from terrestrial contamination. Its interior was found to contain hydrocarbons, fatty acids and dicarboxylic acids, and pyrimidine heterocycles, but no purines. More than 20 different amino acids (totaling about 50 μg/g stone) were present including some of the common proteinogenic ones. There were also isomers not present in biogenic proteins, such as aminobutyrate, isovaline, and several diamino acids (Meierhenrich et al. 2004). The preponderance of some l-amino acids with 20–30% ee (60% in the case of l-alanine; Engel and Nagy 1982; Engel and Macko 1997) is striking; others exist as racemic mixtures of both d and l forms. Obviously, the inventory of organics in the meteorite does not reflect a biological source but resembles the product spectrum from Miller’s and other simulation experiments.
The presence of chiral molecules in extraterrestrial matter has tempted critics of evolution to shift the origin of our biosphere to other cosmic environments. Nobody denies that matter from space reached the planet and mixed with local Earth products which had accumulated in warm little ponds or elsewhere. Thinking in quantitative terms and of the somewhat erratic composition just mentioned, it is unlikely that stone meteorites could have brought in and scattered enantiomeric “seeds” either in bulk amounts or of a defined composition. As the physical and chemical laws of stereoisomer discrimination must have been in force everywhere in the solar system, it should remain our prime interest to understand their effects on Earth.
Fundamental physical discrimination mechanisms
The problem with these scenarios is that the asymmetry effects, calculated ab initio or observed in laboratory simulations, are minute for small molecules such as amino acids and sugars. For example, PVED stabilization of l-amino acids in solution is only of the order of −10−16 kJ mol−1, which corresponds to an excess of 107 molecules of the more stable enantiomers per mole of the racemate (Laerdahl et al. 2000). Nevertheless, it has been estimated that selection towards homo-chiral amino acids and peptides could have been possible in an otherwise undisturbed lake of 1 km × 1 km × 4 m volume during 15,000 years (Tranter 1985). The solution to this problem still requires more theoretical and experimental approaches.
Reaction-induced and phase transition-induced enantiomeric enhancement
Much larger PVEDs and powerful mechanisms for chirality amplification are expected where crystals, polymers, and other macromolecules are involved, and indeed, such systems continue to emerge. In one illustrative example (Blair and Bonner 1981), a mixture of l-leucine and d-leucine with 31% initial l-ee was partially polymerized, and the resulting polypeptide then partially hydrolyzed; the l-ee in recovered unhydrolyzed polyleucine had increased to 55%. Repetition of such partial synthetic and degradative phases, probably during fluctuating environmental dry and wet cycles, were likely conditions on the primitive Earth. In a recent study, when amino acids of low (1–5%) initial l-ee were subjected to slow partial sublimation for several hours, an average fourfold enrichment of the l-forms was observed in the sublimates (Fletcher et al. 2007). This effect can be explained in terms of different vapor pressures of a solid d,l-racemate (reduced due to stronger intermolecular interactions) and the pure enantiomers, just as their melting points and solubilities differ. Sublimation could well have provided an important enrichment mechanism on the way from low ee products, generated by any of the previously described mechanisms, towards the homochirality required in amino acid polymers of high activity and specificity.
Adsorption on asymmetric surfaces
Polar organic molecules adsorb readily to polar mineral phases. If the latter are asymmetric, differential interactions (binding, catalytic, etc.) with enantiomeric substrates and/or products should ensue. Indeed, stereospecific binding of natural l-amino acids and d-glucose, but not d-amino acids or l-glucose, to clay silicates has been observed (Bondy and Harrington 1979). Crystals of common calcite (CaCO3) also display differential absorption of d-aspartic acid and l-aspartic acid molecules on mirror-related crystal growth surfaces (Hazen et al. 2001). This effect is thought to serve as a model for the condensation of homochiral polypeptides on such crystals. Many more stereoselective interactions between organic enantiomers and inorganic mineral supports are likely to exist in nature; these may have contributed to the handedness of chemical evolution from the beginning, and further examples should be sought.
Chirality, conformation, and selection of sugars
Sugars are the other major class of chiral natural products in which asymmetric carbon atoms and the resulting three-dimensional structures of monosaccharides, disaccharides, and polysaccharides determine biological activity. d-Glucose, present in sucrose, starch, cellulose, etc., and d-ribose in RNA are the most prominent sugar compounds. (The nomenclature and apparent contrast of l-amino acids and d-sugars is due to historic definitions and not a basic discrepancy; indeed, both can be derived from the same chiral progenitor d-glyceraldehyde.) The preference of d-sugars has been traced back to parity-violating weak interactions (Tranter 1986), as outlined for enantiomeric amino acids.
An intriguing chemical mechanism of chiral induction was demonstrated (Pizzarello and Weber 2004) when chiral l-amino acids (such as the ones found in meteorites) were used as basic catalysts of the formaldehyde–formose condensation reaction described above and induced 10% ee with “correct” asymmetry in a C4 sugar product (i.e., d-threose).
In nucleic acids, the pentose (C5) sugar d-ribose assumes the furanose conformation which permits easy linkage of mononucleotides via their 3′ and 5′ hydroxyl groups to long 3′–5′-phosphodiester-linked chains and base-paired duplex strands. In RNA and DNA, such polynucleotide structures are stable yet allow a certain amount of flexibility and bending during replication, transcription, and translation. In an impressive series of chemical studies, Eschenmoser and coworkers synthesized pyranosyl-RNA (Krishnamurthy et al. 1996) and DNA homologues with d-glucose and hexopyranose components (Eschenmoser and Dobler 1992; Otting et al. 1993). Not surprisingly, all of these polynucleotides can form base-paired double strands. However, they differ from native RNA and DNA in backbone conformations and exhibit stronger and more rigid base pairing that is apparently less favorable for genetic functions than nucleic acid architectures based on furanosylpentoses. Perfection of the genetic material was achieved with the provision of d-2′-deoxyribotides (see below) that resulted in right-handed double-helical B-DNA of optimum structure and stability. None of the other double-helical DNA structures (A-DNA and Z-DNA, formed in nonphysiological media) has been selected under native conditions.
Heterocyclic bases and nucleic acids
Strangely enough, selection of the same four specific bases which allow and determine stable genetic information also introduced the possibility of spontaneous mutations. For example, hydrolysis of the amino group of cytosine to an oxygen function produces uracil, which would lead to a U=A pair in place of the original C≡G pair during RNA or DNA replication. Such deamination of cytosine bases by chance occurs at low rates (i.e., one event per 108 cytosine residues and round of replication), but it is inevitable. Many of these events will be unfavorable or even lethal; however, in other cases, the modification may produce mutated, functionally improved genes and gene products and, thus, benefit molecular evolution.
Proteins constitute the majority of biological matter and enable a wide variety of cellular functions. The necessity for synthesis of functional polyamino acids from just one kind of stereoisomeric monomer has been established above. A further stipulation is that, not only the genetically defined sequence, but also the macromolecular topography of a polypeptide should be reproducible and predetermined by intramolecular interactions. This requirement of living cells for a stable protein conformation is met by the same universal forces that govern the specificity of nucleic acids. Hydrogen bonds between the H, N, and O atoms of a peptide bond (–CO–NH–; Fig. 7) determine the spontaneous folding of l-amino acid chains into the stable α-helix and β-sheet secondary structures of proteins. The presence of 20 different amino acid side chains attached to a polypeptide backbone allows additional tertiary interactions (not outlined here) that contribute to individual protein stability, folding, and functions. For example, proteins embedded in hydrophobic membranes require specific arrangement of amino acids with hydrophobic residues, while polar side chains are required for catalytic effects, etc. The thermal stability of native structures (inter alia, important for thermophilic organisms) depends on the ratio between the highly ordered and stable and more loosely arranged flexible amino acid sequences. The 20 proteinogenic amino acids fulfill all these chemical demands in an average microbial cell (Weber and Miller 1981), and they were among those available from prebiotic syntheses. Thus, by “natural selection,” they became part of the genetic (ur)code (Table 3).
Chance and necessity
Darwinian selection of individual molecules, contained in a large array of congeners and isomers, that are “fit” for specific binding, function, stability, or any other advantageous property is now being practiced on a large scale in high-tech laboratories which employ combinatorial chemistry for rapid, unbiased development of drugs or other desired compounds. Widely differing chemicals are allowed to react in automated evolution machines with thousands of reaction chambers and are analyzed and selected in high-throughput screening systems. One could, in fact, regard this modern kind of molecular design as a rapid-motion picture of the chemical and biological evolution on early Earth, which also selected the “fittest” molecules and cells albeit over thousands millions of years. Such a comparison is not totally foolish. It confirms the principle that numerous diverse reaction sequences, initially giving rise to random products, can and will, after an undetermined period of time, produce unique molecular systems or organisms fit for a particular function.
The logic of such apparently contradictory events and outcomes was substantiated decades ago by Jacques Monod, French biochemist and philosopher in his famous reflections on “Chance and Necessity” (1970). Years of fierce debate followed about the role of pure chance in life. This was disliked by many, especially when related to human life, but it is undoubtedly in action in gene mutations. The matter need not be reopened here; its biological aspects have been addressed at length by Carroll (2001).
Assembly of supramolecular structures
The paradox is easily explained. As biochemical and biological reactions occur in water, the entropy of water molecules, present in large excess around polymer biomolecules, has to be taken into account. The entropy (∆S) of such a system as a whole (i.e., ordered biomolecules with one hydrate shell, ∆S < 0, plus many disordered water molecules liberated from the monomers, ∆S ≫ 0) will indeed increase over an initial situation in which each monomer molecule had its own water shell. The necessity of including the entropy change of water when assessing the possibility that a complex orderly situation such as metabolism, energy generation, macromolecular structure, or integrity of cell membranes arises spontaneously from the individual components is frequently overlooked.
Polycondensation of amino acids constituted during evolution and still provides today nature’s chemical reactor in which new molecules are created for biological functions. Whenever specific sequences and special three-dimensional structures happen to assemble in random polypeptides (such as synthetic “proteinoids”), these have a certain probability of expressing enzyme-like catalytic activity, inhibitory or regulatory function, metal-binding or other properties.
Problems faced by the chemist when trying to form peptide bonds in aqueous media are that (1) water has to be eliminated from the reactants and (2) amino acids exist in their stable zwitterionic forms +NH3–CH(–R)–COO− (R– are side chains with additional hydrophobic or hydrophilic groups). Expelling water by dry heat usually leads to heterogeneous products and damages chemically sensitive residues R and reduce their biochemical potential. Nevertheless, a variety of thermal polyamino acids has been prepared and studied as “metabolic microsphere” models of protocells (Rohlfing 1976; Fox 1980). Specific amino acid sequences may be enriched in such products (Hartmann et al. 1981). More gentle reaction conditions prevail in the presence of carbonyl sulfide, a simple volcanic gas, which brings about formation of dipeptides and tripeptides by “polymerization on the rock” at moderate temperatures (Leman et al. 2004).
All these attempts to abiotically produce polypeptides suffer from uncertainties such as availability of monomers, suitable reaction conditions on early Earth, and poorly defined product structures. In the early 1990s, Rode and coworkers (Rode and Schwendinger 1990; Schwendinger and Rode 1992) tackled the lingering problem of prebiotic amino acid condensation with a new type of chemistry, salt-induced peptide formation or “SIPF.” This has become a most productive approach and now represents the simplest way to oligomerize amino acids in aqueous solution (Rode 1999). Peptide bonds form under the dehydrating action of high (≥3 M) NaCl concentrations in the presence of copper(II) ions, which serve to complex and coordinate educt and product molecules. To mimic early Earth and ocean conditions realistically, it is possible to start from dilute solutions which are then subjected to evaporation cycles. An earlier short study (Okihana and Egami 1979) of amino acid oligomers produced in hot sea water enriched with transition metals may have employed a similar mechanism.
The SIPF reaction produces peptides up to the heptamer stage within a few days but long-term reactions have not been described. When the reaction is combined with adsorption on clay minerals, chain elongation to higher peptides is favored and hydrolytic backreactions are reduced. While the amino acids in general condense to all possible homodipeptides and heterodipeptides, it is an important feature that certain sequences are preferred, e.g., histidine plus hydrophobic residues, and others are under-represented. It is of considerable interest that the dipeptides preferred in SIPF chemistry appear to be reflected in dipeptide frequencies found in membrane proteins of ancient archaebacteria and other prokaryotic cells; this may not be a fortuitous correlation.
Finally, it is overdue and rewarding that attention is now being paid to the rare, yet physiologically important, amino acids histidine, methionine, and tryptophan in peptide formation and chiral discrimination (Reiner et al. 2006). These protein constituents have been neglected for a long time, mostly because they are chemically less stable than other amino acids and also because they were late additions to the genetic code. Histidine deserves interest for its outstanding role in enzyme catalysis and methionine as the N-terminal starter amino acid in ribosomal protein synthesis.
The present knowledge and potential of abiotic amino acid and peptide chemistry have led Rode and coworkers to proclaim an early peptide–protein world that allowed all essential functions of life such as self-organization and growth (Rode 1999). Support for this view has come from the case of a self-replicating α-helical peptide (Lee et al. 1996) which combines several of the preferred dipeptide sequences mentioned above in its 32-amino-acid sequence. A final answer to the sibyllic reflection “Can even peptides do it?” accompanying Lee’s communication in Nature is still not known.
Both polypeptides and polynucleotides arise by elimination of water between the component monomers. The chemistry of assembling a 3′–5′-linked polyribonucleotide, however, differs from peptide bond formation in that negatively charged phosphodiester bonds –O–PO2−–O– are more reactive than a neutral, stable peptide bond –CO–NH–. Thus, metal ions that neutralize the negative charge and the adjacent 2′-hydroxyl groups of ribose can enable nucleolytic backreactions. (This is, inter alia, why DNA lacking that group was able to become a second genetic molecule alongside RNA, despite the extra investment in enzyme catalysis.) Moreover, in condensation reactions, the 2′-OH function competes with the formation of the regular 3′–5′ internucleotide bonds. On early Earth, the isomer problems could have been reduced under plausible fluctuating (hot and cold, dry and wet) environmental conditions in which various random copolymers with mixed 2′–5′ and 3′–5′ bonds formed, but the more thermostable double-helical strands enriched in 3′–5′ bonds survived subsequent hydrolysis and accumulated (Usher 1977). As zinc ions also favor 3′–5′-bond formation (Lohrmann et al. 1980), it is probably not fortuitous that modern RNA and DNA polymerases are zinc metalloenzymes.
Nevertheless, generally applicable conditions for abiotic nucleic acid formation are still not readily available. Recent results (Mansy et al. 2008; Rajamani et al. 2008) of “lipid-assisted” ribonucleotide polymerization were unanticipated previously and may mark a breakthrough. When mixtures of nonactivated mononucleotides (AMP or UMP) and phospholipids were subjected to a series of hydration–dehydration cycles at 90°C, RNA-like polymers of 50 or higher chain length enclosed in lipid vesicles were produced. Other vesicles with “leaky” membranes made of fatty acid and fatty alcohol glycerides (without phospholipids) were able to retain encapsulated oligonucleotides but allowed the passage of nucleotides from the outside and efficient template copying in the interior.
Not all details of abiotic polynucleotide formation can be satisfactorily explained at this time. Therefore, a search continues for chemical alternatives that possibly predated RNA and might have combined simpler backbone structures with the firmly established hydrogen-bonded base-pairing scheme of genetics. Problems inherent in an RNA world (see below) could also have been alleviated by such a precursor. To this aim, ribose-3′–5′-phosphodiester links have been replaced by threose (i.e., a C4 sugar) or even glycol (C2) and by peptide links (Schoning et al. 2000; Zhang et al. 2005; Nelson et al. 2000). While interesting synthetic compounds were obtained, it seems unlikely that these “TNA,” “GNA,” and “PNA” nucleic acid analogs were relevant in chemical evolution. There is no real evidence for their prebiotic occurrence, and the necessary later transitions to present-day molecular biology would have created new, probably insurmountable problems.
Ribosomes and peptide synthesis
Bacterial ribosomes are small particles, composed of three different ribosomal RNAs (rRNA) and more than 50 ribosomal proteins, on which new proteins are synthesized. These are assembled from activated amino acids (aminoacyl adenylates and aminoacyl-tRNAs, –CO–OPO2–O–AMP, –CO–O–tRNA) in the sequence programmed by messenger RNAs (mRNA), which were transcribed from genes. The various steps of ribosomal protein synthesis, which is universal and highly conserved, are not detailed here. What is important is that the intracellular assembly of ribosomes from their numerous constituent molecules is spontaneous. It is also amazing that, under suitable ionic conditions, reconstitution can proceed in vitro (Nomura and Erdmann 1970; Nissen et al. 2000). In a very complex yet predominantly error-free process, ribosomal proteins bind to each other and to the rRNAs, which serve as scaffold, in an orderly fashion. The particle is held together by numerous hydrogen bond and ionic interactions. The ease and fidelity of ribosome assembly suggests that it was an early invention in the prebiotic era; this new mechanism then overcame the limits of thermal or SIPF. The fact that ribosomes and their peptidyl transferase activity may be considered primordial ribozymes (Nissen et al. 2000) also points to an early origin, dating back to the putative RNA world.
It is appropriate to mention that a seemingly simpler, energy-saving, nonribosomal mechanism of peptide formation exists (Kleinkauf et al. 1992). In this system, amino acids activated by ATP are transferred and linked to an enzyme complex on which a peptide chain grows via condensation of thioester (–CO–S–) functions. Such a transthiolation–thiotemplate pathway could have neatly matched and served an iron–sulphur world. However, the present mechanism is restricted to small peptides of special functions and not suitable to generate larger enzyme or fibrous protein structures. Whether both pathways may have operated in parallel in the beginning remains a matter of speculation.
Enclosure of organic matter, be it a watery soup or sedimented particles, inside lipid membranes or vesicles is straightforward. Entropy-driven assembly of monomolecular surface layers and double membrane-coated vesicles or liposomes (Fig. 8), achieved by spreading and agitating lipids upon and in aqueous media, are standard experiments in physical chemistry teaching laboratories. Phospholipids were obtained from glycerol, phosphate, and hydrocarbon derivatives during evaporation on a hot sand or clay matrix (Hargreaves et al. 1977) and indeed formed the expected bilayer membranes. The advantage of lipid vesicles for synthesis of RNA-like polymers (Mansy et al. 2008; Rajamani et al. 2008) has been described above. Entrapment of native nucleic acids in artificial liposomes (Monnard et al. 1997) is also remarkably easy. Lipid membranes themselves integrate hydrophobic proteins (i.e., composed of ancient hydrophobic amino acids such as leucine, valine, phenylalanine; cf. Table 3) as channels or receptors for communication with the aqueous world outside. Current expertise in liposome and protein biochemistry combined has facilitated progress in simulating functional gene expression networks (Ishikawa et al. 2004).
Cell evolution and membrane topology including a reference to warm little pond chemistry have been discussed recently in context (Griffiths 2007). A “living” protocell requires that many, if not all low-molecular-weight and high-molecular-weight compounds are encapsulated within a membrane-coated vesicle. Experiments to assemble such a body have not yet been tackled and may have to await a future LUCA Revival Project as prophesied below. Meanwhile, the demands for synthetic protocell membranes are obvious. Different lipids such as glycerol esters, fatty acids, and fatty alcohols of various chain lengths and double-bond pattern must be combined to permit structural heterogeneity and flexibility (rather than rigidity) and a certain degree of leakiness (rather than impermeability). The art of designing tailor-made membranes for protocells has hardly begun.
Molecular machinery of life: metabolism, energy, catalysis, and a genetic apparatus
Living cells require energy and nutrients to sustain their bare existence and to generate daughter cells by division. Long before chlorophyll-like pigments and phototrophic membranes were present and allowed the harvesting of light as an energy source, early anaerobic protocells could only vegetate by chemolithotrophic forms of metabolism. Ammonia, hydrogen, nitrite, sulfide, and the iron–sulphur system could serve as reductants, and C1, C2, and C3 compounds like methane, methanol, formaldehyde and formate, ethanol and acetate, glyceraldehyde and amino acids taken up from the outside as substrates of an anabolic metabolism. In our opinion, it makes little sense to relate such cells with any kind of extant chemotrophic bacteria (e.g., methanotrophs or methylotrophs) all of which are much more specialized today. However, what can be safely assumed, from the great metabolic diversity of modern microbial life, is that some chemically and thermodynamically feasible reaction sequence (unknown in detail) was established that was able to produce, inter alia, ATP via substrate chain phosphorylation for intracellular energy. Formation of some group-transferring coenzymes such as ADP-glucose (Mar and Oró 1991) and pantetheine/coenzyme A (Keefe et al. 1995a, b) under primitive conditions has been reported. During a phase of random assembly of both synthetic and lytic pathways, rudimentary catalytic activities would have developed, and these catalysts would have been protected against breakdown when saturated with substrates (de Duve 1987, 1991). An increasingly productive network of metabolic extension and catalytic innovation could thus have been established. Was metabolism first?
Inventories of genes required for extant minimal bacteria (Mushegian and Koonin 1996) and corrected for the earliest ancestor, LUCA
Minimum number in extant bacteria
Not required in LUCA (cf. text)
Basic genes/gene productsa required for LUCA (incomplete list)
Translation (ribosomal protein synthesis)
Some (<20 amino acids present)
Ribosomal RNAs and proteins
Replication (nucleic acids)
Kinases, polymerases, ligases
Transcription (DNA → RNA)
Majority (no DNA present)
Chaperones (protein folding)
Many (no deoxynucleotides)
Kinases, ribosyl transferases
Amino acid metabolism
Some (fewer biosyntheses)
Nonribosomal peptide synthesis
Thiol peptidyl transferase(s)
Fatty acid synthetases
Energy production (aerobic)
Many (no oxidases, etc.)
Energy production (anaerobic)
ATP via substrate chain phosphorylation
Aldolases, glycosyl transferases
Uptake of inorganics (metals)
Chelating peptides, thionins, thioredoxins
Uncertain (leaky membranes?)
It is worth noting that the enormous variety of presently known enzymes depends on a much smaller number of principles when it comes to specificities and catalytic mechanisms. The capacity of acidic or basic amino acid side chains (aspartate, glutamate; arginine, histidine, or lysine) to promote acid or base catalyses is obvious. Where substrates and products differ only in the position and number of a few hydrogens, catalysts can produce intramolecularly rearranged isomers and dehydrogenated (=oxidized) or hydrogenated (=reduced) products. Shifting hydrogens plus electrons between carbon or nitrogen atoms in such cases is a common and facile capacity at enzyme active sites as the smallest atoms, hydrogens, are easy to translocate. Another frequent and probably very ancient structure motif is the “nucleotide fold” of dehydrogenases, kinases, and other related proteins which determines their substrate specificity for nucleotides, ATP, and several coenzymes (Rossmann et al. 1974). All these basic principles, optimized in later modification and selection steps up to present-day enzymes, were certainly of great advantage in shaping early metabolism.
The genetic code
The molecular and informational bases of heredity are the nucleotide sequences in DNA genes and RNA transcripts. These carry the genetic code for their translation into the amino acid sequences of enzymes and all other cellular proteins. Initially excluding DNA (see below) but with RNA strands and abiotic amino acids at hand, an ur-code must have been one of the earliest and most important steps in the evolution of life. The rationale and necessity for a degenerate three-letter code of four nucleic bases (A, C, G, U) arranged in 43 = 64 codons that specify up to 20 (or 21, including selenocysteine) different amino acids plus stop signals, present in all modern organisms, is textbook knowledge. The code is reproduced and discussed in a recent article in this journal (Penzlin 2009). It started with few primeval amino acids (i.e., those listed in Table 3) and subsequently expanded when further metabolic descendants became available, in a fashion that has convincingly been explained by the coevolution theory of Wong (1975). For example, of eight original serine codons UCx and UGx (x = any base), the latter four were used for coding cysteine and tryptophan, respectively, when their biosynthesis pathways from serine became established. In contrast, a possible doublet code with only 16 places would have lacked the potential for such expansions and died out if it had existed as an early competitor.
RNA and ribozymes first?
Ribonucleic acids are complicated and versatile molecules. With an anionic phosphate, hydroxyl groups, hydrophobic heterocyclic bases, basic nitrogen atoms, and hydrogen bonds, they combine different chemical structures and functions, with the notable exception of redox reactions. Their abiotic formation, selection from isomer mixtures, and roles in ribosomal protein biosynthesis and the genetic code have been described above. A conspicuous property of oligonucleotides and polynucleotides is the capacity to engage in intramolecular hydrogen bonding and to fold into many specific three-dimensional structures (Fig. 9) that enable different functions. At ambient temperatures, three or four adjacent base pairs (i.e., six to eight hydrogen bonds) suffice for engagement in specific, reversible interactions such as codon–anticodon recognition. Twenty or more base pairs produce thermostable folded structures as in transfer-RNAs and ribozymes. Such structures and, consequently, biological functions arise and exist solely from a defined chemical situation (nucleotide sequence) and the physical conditions.
A boost for evolution research came with the discovery that certain RNA molecules act like enzyme proteins in that they accelerate chemical reactions without themselves being changed in the process (Cech 1987). An “RNA world” of early evolution was proclaimed in which such “ribozymes” serve to catalyze the synthesis of new RNAs and other products (Gilbert 1986). Ribozyme activities have been described in increasing number (Joyce 1998), and the third edition of “The RNA World” (Gesteland et al. 2006) now combines contributions from 66 authors in a volume of 750 pages. Nevertheless, questions like “Did life really start out in an RNA world?” (Waldrop 1989) or “RNA seeks its maker” (Piccirilli 1995) remain and are justified by some serious objections.
Ribozymes such as the self-splicing intron drawn in two dimensions in Fig. 9 are single-stranded RNA molecules, at least 50–60 nucleotides in length, in which short double helices and loops are crowded together in three dimensions maximizing hydrogen bonding. At an exposed site, phosphodiester bonds can be broken (hydrolyzed) or rearranged (transesterified) between ribozyme and RNA substrates, or ribozyme and amide or peptide substrates. Such reactions have clearly been useful in cutting, joining, and moving pieces of RNA in the course of creating novel polynucleotides and polypeptides with new functions. A reaction sequence in which pyrimidine nucleotides are put together from a ribose phosphate and a base and added onto an RNA substrate (Unrau and Bartel 1998) is of particular interest. In contrast, reactions like ribozyme catalysis of Diels–Alder cycloadditions observed in vitro have no physiological counterpart. Redox reactions like the action of dehydrogenases, which are crucial for cellular energy and anabolic metabolism are hard to imagine in an RNA world lacking enzymes and coenzymes. Finally, the most serious and puzzling question is the chicken-and-egg problem of how first catalytic RNAs came into existence themselves. Claims that RNA was preceded by another class of abiotic self-replicating molecules, probably peptide nucleic acids (PNA; Böhler et al. 1995; Nelson et al. 2000), in which ribose–phosphate links are replaced by amide or peptide bonds do not help in unraveling the problem. The contribution of ribozymes to primordial chemical evolution still requires evaluation.
DNA: the stable genetic information
Once DNA together with DNA-replicating enzymes had assumed the role as a hydrolysis-resistant carrier of genetic information, the path of evolution opened towards higher complexity and diversity (Carroll 2001). Ribo-organisms like the last universal ancestor LUCA were succeeded by Archaea and Bacteria with their DNA genomes and the nucleated cells of Eukarya with an even higher DNA content. It has been estimated that the potential of DNA genomes is far greater than is actually realized in organisms. Repetitive DNA, nontranscribed introns, “selfish” (presumably useless) DNA in higher organisms (Orgel and Crick 1980), and other peculiarities of extant cells mean a waste of energy and scarce resources when looked at from the very modest beginnings. However, they provided a basis for continuing gene duplications and more complex forms of life. Should the origin of DNA precursors, ribonucleotide reduction, be given yet another “world” status?
Metabolism, protein, RNA, lipid, iron–sulphur world: ur-world prophets unite!
What was LUCA, and how did it live?
The last universal common ancestor (LUCA) from which all current living organisms are descended is a hypothetical protocell that arose at an unknown point of time more than 3,500 mya. We are undecided whether to think of LUCA as a genetically defined individual species or a small commune sharing genetic information by frequent horizontal gene transfer (Woese 1998; Whitfield 2004). The division in time and in complexity between LUCA and the most ancient fossil microorganisms, which are more than 3,000 million years of age and abound in Precambrian sediments in several areas of the world (Pflug and Jaeschke-Boyer 1979; Schopf 1993) is also uncertain.
It has been reasoned (Lazcano and Miller 1994) that once cellular life had started the time taken for a transition to autotrophic, photosynthetic cells may have been relatively short in the evolutionary timescale—probably less than ten million years. Many metabolic mechanisms and enzymes exhibit considerable analogies and similarity in the different life forms of today; thus, new genes required for photophosphorylation, CO2 fixation, etc. might have been generated from tried gene sequences of the ancient chemotrophic pathways by straight gene duplication events.
Some characteristics of LUCA and its environment have recently been deduced from the G+C content of ribosomal RNAs (rRNA). This base composition is strongly correlated with the optimal growth temperature of prokaryotes due to the higher stability of G≡C base pairs (Galtier et al. 1999; Boussau et al. 2008). Ribosomal RNAs and protein sequences from hundreds of modern species have been analyzed as “thermometers” for backward extrapolation, and these suggest that LUCA most likely still had an RNA genome. (In possession of proteins, however, it cannot have been “the last ribo-organism” cited by Benner and Ellington 1987). It strived for life in anoxic waters at about 60°C, well before thermophilic Bacteria and hyperthermophilic Archaea sprouted from the big tree (Fig. 10). These ancestors of all modern organisms (including ourselves, the Eukarya) may have had to survive a later, hot period in early oceans (≥80°C) after a big impact event to which they responded with the acquisition of more thermostable DNA genes.
LUCA must have possessed a minimum set of genetic elements to code for its metabolic demands and reproduction. Such a set was deduced from the genomes of two very small extant bacteria (Mycoplasma genitalium and Haemophilus influenzae) and it includes 256 genes (Mushegian and Koonin 1996). LUCA should have managed with even fewer genes if it lacked, inter alia, redox enzymes for aerobic respiration, DNA synthesis, and complex amino acids not yet fixed in the ur-code. Our provisional guess is that no more than 200 genes (Table 4) were sufficient for its existence. This estimate is consistent with earlier figures arrived at in a different context and using different approaches (Gil et al. 2004). Such considerations certainly do not imply that LUCA and present-day parasitic microorganisms have anything in common except a very small number of genes.
Can LUCA be reconstructed in the test tube along these lines and revived? Success is not guaranteed but experiments are worth the effort. A promising strategy may be to follow the recent protocols (Gaucher et al. 2003, 2008) to restore just one kind of primeval protein (i.e., ribosomal elongation factor Tu). Retrogression from protein and gene sequences of numerous extant microorganisms near the roots of the big tree towards LUCA’s putative genes and gene products would be carried out by computational analyses. From these data, synthetic gene construction, protein expression, and activity measurements might eventually provide the minimum number of catalysts and the genetic elements required for anaerobic, chemotrophic life. Their molecular interactions could be analyzed by modern genomics and proteomics techniques and potentially adverse combinations could be deleted or modified. Inclusion in ready-made lipid vesicles (Fig. 8) spiked with hydrophilic channels might facilitate the project. Incubation in defined artificial media will have to include a wide variety of inorganic and organic potential nutrients, proceed at different (mesophilic? thermophilic?) temperatures, and be monitored by the most sophisticated analytical tools available in chemistry and microbiology.
Oh what a big If : Will, at some point of the LUCA Revival Project, the contents of some culture bottle (“a warm little pond… with all sorts of salts… heat etc. present”) sustain a metabolism and even be able to multiply for a little while? Oh what a wonderful Yes! How, then, fellow biologists, do we name the new resurrected species of ancestor or ancestress: Luca resurrectus or resurrecta?
What a big If : Certainty out of variety
It is often said that all the experiments and facts described above do not and, indeed, cannot prove that cells capable of energy production, anabolic metabolism, and genetic reproduction (i.e., present-day life) could ever have arisen from inanimate chemical matter. Such proof is compounded by the extreme complexity of living systems on one hand, and gaps and inconsistencies that remain in our knowledge of prebiotic chemistry and self-assembly on the other hand. Low yields of sugar formation, poor catalytic power of random polypeptides, and the tiny probabilities calculated for the emergence of large functional polymers are among the purported problems. Such objections, repeated again and again by creationists and other skeptics are fading and will continue to fade as more and more reasonable synthetic pathways, geochemically plausible conditions, and thermodynamic details of molecular interactions become known. Understanding spontaneous structure formation even in large, oligomeric biomolecules such as the ribosome or membrane-integrated electron transport chains have been more recent major advances. Nevertheless, there are still basic problems and misconceptions in our thinking about molecular and cellular evolution that need to be addressed.
Has there been enough time?
Humans are used to thinking about the past in years, centuries, and millennia and possibly even a million years when confronted with geological and paleontological records. Time spans of several thousand million (109) years, such as the age of our universe and the solar planets including Earth, are beyond our imagination. Therefore, it is difficult for many to conceive that astrophysical and chemical processes which are very slow on our human timescale have a major impact on atmospheric and terrestrial environments. These can create adverse or, in other situations, favorable conditions over very long periods of time (i.e., hundreds of million years) that are sufficient to produce ever more complex molecules stable enough to accumulate and eventually generate cell-like, growing, and multiplying vesicles.
A more precise, statistical argument has been that the age of the Earth was simply not sufficient to allow assembly of oligomeric and polymeric biomolecules from the appropriate monomers by chance when the huge number of all possible combinations is considered. Experience among protein chemists and recent progress in prebiotic peptide research (Rode 1999) show that specific monomer alignments are intrinsically preferred, thus reducing the numbers of possible combinations and significantly weakening the argument. Insufficient time for molecular evolution on planet Earth may indeed be a pseudoproblem.
How likely, or unlikely, is the assembly of a cell “from scratch”?
Critics of chemical and cellular evolution sometimes resort to probability calculations. They hope to rule out the aggregation of a self-reproducing cellular body, subcellular fraction, or large molecules without support from unknown, supernatural forces. For example, in a popular German science magazine (1982), the chances for a scenario in which “correct” macromolecules like genes assemble de novo from a collection of atoms or small monomers were figured to be as low as 1:11,200,000. Not surprisingly, the author concluded that such an origin of life was statistically impossible.
Naïve approaches like this neglect well-established facts. Complex biomolecules, large or small, are built in a modular fashion. Smaller, submolecular structures and functional domains originate separately and then combine, as in the following examples: A heterocyclic base and the sugar phosphate moiety of a nucleotide come together ready-made from different pathways. In proteins, stable subdomains such as fibrous α-helical stretches, globular β-barrel structures, hydrophobic transmembrane regions, or sites specific for binding coenzymes, carbohydrates, or enzyme substrates can be clearly distinguished. Functionally and structurally separate subsections are found in the cloverleaf structure of tRNAs (the amino acid-accepting and the anticodon stem), in ribozymes, and in DNA strands (“genes in pieces”). Many enzymes specifically incorporate magnesium, calcium, iron, zinc, or other metal ions after their synthesis to reach their final activity or stability. Membranes assemble spontaneously due to hydrophobic interactions.
It is obvious that assembling macromolecules from such smaller preformed components in solution or inside the small volume of a lipid vesicle increases the overall probability of establishing favorable new properties in a population of organic molecules by many orders of magnitude. It is also obvious that a sound knowledge of extant biomolecules, including their syntheses, structures, and functions, is required to make reasonable predictions regarding the origin and properties of probable prebiotic progenitors. It seems clear that objections raised by critics who lack or deny such background knowledge can be disregarded.
This is an extensive assertion. Therefore, let us illustrate and quantify the situation with one more specific yet central problem in a statistically proper way. How could cell-like compartments acquire genetic systems made of uniformly configurated and linked monomers if the soup or slime of abiotically formed constituents contained racemic mixtures of ribose and other sugars, heterocyclic bases unsuitable for base pairing, and “wrong” internucleotide bonds (Kuhn 1972)? If at each single position of a growing oligomer chain the probability of incorporation of a “correct” d-ribose, 3′–5′ linkage, and a proper base was only 1/100, then it would be (1/100)10 = 10−20 for a uniform, “useful” decanucleotide and 10−100 for a 50-mer (the size of a small tRNA). That a molecule of the latter kind would ever form by chance is indeed unrealistic. However, the first probability is not too small to be feasible as Avogadro’s number says that 1 mmol of random decanucleotides (about 3 g of abiotic matter) has 6 × 1020 molecules. Thus, the “unlikely” formation of a correctly structured, base-paired decanucleotide fit for replication or another function (Fig. 9) must occur and generate new potential. Evolution proceeded in many small steps, in parallel or succeeding each other. Students interested in both biological and physical chemistry can simulate and verify this claim on their computers at any time (Försterling et al. 1972; Kuhn et al. 2009). Critics should familiarize themselves with such straightforward ways of thinking, too.
Will we ever and, indeed, do we have to know the historical path of chemical evolution? This may be the most interesting and fascinating question in evolution research. We wish to satisfy our inborn sense of curiosity and our desire for universal knowledge of this world. However, surprisingly, the answer to both questions is “No.”
The first “No” is obvious. Progress will be made, but the heritage of 109 years of the most complex abiotic, biogenic, and biological chemistry on Earth is not amenable to complete unraveling by us now or in subsequent generations. Even with the very best sets of experimental data available, nobody will be able to define which of innumerable possible steps of synthesis and decay and chaotic and stagnant moments really happened. It is also doubtful whether machines can ever do it.
The second “No” is of more interest. It should be possible someday to identify at least one model sequence of events, hypothetical but supported by several lines of solid evidence, which rigorously satisfies all molecular, thermodynamic, and kinetic constraints and requirements. We would still not know the historical path but no one could continue to argue that an evolution of life driven by chance and necessity was impossible for objective reasons. Considering the multiple, redundant physical and molecular possibilities now known and described above—not by far comprehensively covered in our review—the authors are certain that we are approaching the stage of enabling transitions from inanimate matter to a replicating entity inside artificial vesicles. Computer studies of intermolecular networks “in silico” have their place, but bright new ideas for intelligent experiments in vitro and in vivo are, at the very least, equally rewarding.
In this review, we have listed around a hundred critical steps, reactions, and consequences that must be analyzed and combined in order to proceed from speculation to an increasingly rational understanding of the chemical evolution that occurred very early in the history of our planet. The dictate of physics and chemistry over molecular aggregates and cells did not end after LUCA had started to proliferate, mutate, and “change the world.” They continued to exert their influence on energy metabolism and reproduction during the transition from anoxic, reducing conditions to an oxidizing environment when light absorption by chlorophylls and oxygenic photosynthesis (splitting of water to protons, electrons, and molecular oxygen) became established about 2,000 mya. Photosystems and oxygen permitted efficient energy production via electron and proton potentials in specialized intracellular membranes in bacteria, some of which became the energy-providing endosymbionts of eukaryotic cells (Kutschera and Niklas 2004). Multicellular organisms began to step out of the waters and master terrestrial life. These epoch-making events during later eons of the evolution of life on Earth are not the subject of this review. They are in part dealt with in the following papers.
One of the most complex organisms, Homo sapiens, is often highly critical of thinking that he or she should be no more than part of an early chemical and long biological evolution that included incredibly numerous, accidental, and random steps. It is, therefore, still believed by some that supernatural principles have to be cited to satisfy the long-cherished yet irrational desire of man to be unique among all other things; the authors do not agree that any are required. As biochemical and biophysical research and reasoning begin to understand even extremely complex physiological phenomena, such as consciousness and memory, sexual differentiation, or aging, to name but a few, there is hope for more informed and tolerant relations between us and our biological descent.
Charles Darwin has been admired, scolded, and ridiculed for 150 years. Next to his Origin of Species and Descent of Man, the picture of the “warm little pond,” probably coined incidentally, has become one of the more popular quotations. The rhetorical question “Goodbye to the Warm Little Pond?” asked by Science in 1990 after the advent of the iron–sulphur surface world idea is not only enjoyable reading but is once again a testament to Darwin’s ongoing, amazing intellectual impact on science and culture.
Vivid discussions with colleagues and friends, in particular Horst-Dieter Försterling, Peter Kaiser, Hans Kuhn, Ulrich Kutschera, Stanley Miller (deceased; La Jolla, CA, USA), Klaus-Heinrich Röhm, and Günter Wächtershäuser are greatly appreciated. The students attending summer academies on evolution at La Villa and Alpbach (Tyrol), sponsored by Studienstiftung, contributed many critical questions and stimulating ideas.