Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Corresponding author: N. G. Holm. Tel.: +46 8 164743; fax: +46 8 674 7897; e-mail: email@example.com
Magnesium plays a special role in biochemistry because of its ability to coordinate six oxygen atoms efficiently in its first coordination shell. Such oxygen atoms may be part of one or two charged oxyanions, which means that Mg2+ can, for instance, tie together two different phosphate groups that are located at distance from each other in a macromolecule, and in this way be responsible for the folding of molecules like RNA. This property of Mg2+ also helps the stabilization of diphosphate and triphosphate groups of nucleotides, as well as promoting the condensation of orthophosphate to oligophosphates, like pyrophosphate and trimetaphosphate. Borates, on the other hand, are known to promote the formation of nucleobases and carbohydrates, ribose in particular, which is yet another constituent of nucleotides. The oldest borate minerals that we find on Earth today are magnesium borates. Dissolved borate stabilizes pentose sugars by forming complexes with cis-hydroxyl groups. In the furanose form of ribose, the preferential binding occurs to the 2 and 3 carbon, leaving the 5 carbon free for phosphorylation. The central role of Mg2+ in the function of ribozymes and its ‘archaic’ position in ribosomes, and the fact that magnesium generally has coordination properties different from other cations, suggests that the inorganic chemistry of magnesium had a key position in the first chemical processes leading to the origin and early evolution of life.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Magnesium (Mg) is a common element on Earth and the other terrestrial planets. It is one of the eight main elements of Earth’s crust and one of the four major elements making up the mass of the whole Earth. Furthermore, Mg is one of the principal constituents of silicate minerals that build up Earth, like olivine, pyroxenes, and Mg layer silicates (e.g., serpentines, talc, Mg smectites). The concentration of divalent magnesium (Mg2+) in contemporary ocean water is 52.8 mmol kg−1. The coordination geometry of magnesium is normally octahedral, that is, the Mg atom coordinates six atoms – almost always oxygen – around itself in its first coordination shell (Kehres & Maguire, 2002). Six-coordinated Mg2+ has a small ionic radius of 0.65 Å, at the same time as it has the largest hydrated radius of any common cation. The volume difference between hydrated and ionic Mg2+ is almost 400-fold (Kehres & Maguire, 2002). In the marine geochemical environment, magnesium is particularly important because the tri-octahedral layer of the common smectites in sediments consists primarily of brucite, the mineral name of magnesium hydroxide (Mg(OH)2). The ocean floor beneath the sediment layers consists of basalts and ultramafic rocks that have a high content of primary ferromagnesian silicate minerals (olivine and pyroxenes). Alteration of these minerals in contact with water leads to ‘serpentinization’, a process in which olivine and pyroxenes are transformed to serpentines. Serpentines like lizardite cannot accommodate all of the magnesium of the primary minerals, so brucite is formed as a separate mineral phase, often in veins of the serpentine, at temperatures below about 315 °C (Holm & Neubeck, 2009). Laboratory experiments of olivine alteration by our own research group show a spike of dissolved Mg2+ at about 25 ppm when the olivine surface is fresh and then a strong decrease in the concentration of the fluid phase with nucleation and precipitation of secondary Mg phases, such as brucite (Neubeck, 2011).
Brucite may be transformed into double-layer hydroxides (DLH), also called layered double hydroxides (LDH), if a fraction of the divalent Mg2+ is replaced by common trivalent cations such as Al3+, Fe3+, and Cr3+ (Arrhenius et al., 1997).
Magnesium as a Catalyst
Magnesium ions play critical roles in cellular metabolism. They stabilize structures of proteins, nucleic acids, and cell membranes by binding to the macromolecule’s surface (Yang et al., 2004). They are also key to enzymatic reactions in various ways; for instance, they can generate magnesium-substrate scaffolds to which enzymes bind. However, the best known example of divalent Mg in life processes is probably the central role that it has in photosynthesis as a component of chlorophylls. Anoxygenic photosynthesis occurred by photosynthesizing anaerobic bacteria already before the existence of cyanobacteria (Olson, 2006). Reported evidence suggests that the photosystem in the anoxygenic purple photosynthetic bacteria is the most ancient known (Blankenship, 2001). However, divalent magnesium has many more functions in biological systems; of the many metal cations that form complexes with nucleotides, nucleic acids and enzymes, perhaps none is more essential than Mg2+ (LaRowe & Helgeson, 2006; Freisinger & Sigel, 2007). According to the ‘RNA World’ hypothesis, the first enzymes – the ribozymes – consisted of ribonucleic acid (RNA), which depended on Mg2+ for its self-cleavage (Gilbert, 1986; Cech, 2002). Divalent magnesium is also present in all deoxyribonucleic acid (DNA) and RNA activation processes (Anastassopoulou, 2003). These circumstances point to a central role for Mg in the geochemistry that presumably led to the first life-like processes. Maguire & Cowan (2002) have reviewed the unique properties of Mg2+ among biologically relevant cations. The Mg2+ is a cofactor of enzymes and an essential catalyst for many biochemical reactions (Yamagata et al., 1995). For instance, polyphosphate generation often has an absolute requirement for divalent metal ions, typically Mg(II) (de Zwart et al., 2004). It is also well known that Mg2+ is required for the stabilization of the diphosphate group of adenosine diphosphate (ADP) and the triphosphate group of adenosine triphosphate (ATP) (Lindh, 2005). The reason is that Mg2+ forms six-membered rings with the oxygen and phosphorus atoms of ADP and ATP (Fig. 1). In solution, complex formation with the β- and γ-P of ATP is favored (Lindh, 2005). Divalent magnesium is also known to facilitate the stacking of adenosine monophosphate (AMP), which may promote oligomer formation (Herrero & Terrón, 1998). The intra- and extracellular concentration of Mg2+ in organisms is about 1 mmol kg−1 (Grubbs, 2002; Lindh, 2005). Cytosolic (intracellular fluid) ATP binds up to 50% of total cellular Mg2+ in virtually all cells (Maguire & Cowan, 2002). Divalent magnesium in aqueous solution can form the deprotonated Mg(OH)(aq)+ complex that still binds to nucleotides like RNA (Walter & Engelke, 2002). It has been shown by several investigators that magnesium pyrophosphate (MgPPi) is easily formed under mild abiotic hydrothermal conditions (165–180 °C) from magnesium salts and orthophosphate (Pi) (Seel et al., 1985, 1986; Kongshaug et al., 2000). PPi formation is promoted by low activity of water in the system, for instance through low water-to-rock ratio (Russell & Hall, 1997). Seel and co-workers instead used magnesium monohydrate phosphate dispersed in water (saturated solution) in their syntheses, whereas Kongshaug and co-workers obtained low water activity by the use of phosphoric acid. Hermes-Lima & Vieyra (1992) were successful in condensing pyrophosphate (PPi) from orthophosphate at room temperature in the presence on a fresh precipitate labeled as ‘magnesium phosphate’. To reduce the activity of water, they replaced 80% of the water by dimethyl sulfoxide. Pyrophosphate formation was shown to be most efficient above pH 9. However, because the method that they specified involved high pH, it is likely that this ‘magnesium phosphate’ that they produced was actually brucite or a mix of brucite and protonated Mg-phosphate, in analogy to that produced for scavenging of low concentrations of phosphate in ocean water by the procedure of Karl & Tien (1992).
Condensed Phosphates and Their Role in Biochemistry
A major reason why nature has chosen phosphate for so many purposes in living systems is that it can link up to two adjacent groups, like in nucleic acids, and still remain ionized (Westheimer, 1987). For this purpose, the linking ion has to be at least trivalent. Therefore, phosphate derivatives like pyrophosphate (), complexed to Mg2+, can be a ubiquitous leaving group in biochemistry.
Mulkidjanian et al. (2008) have suggested an earlier transport of Na+ over H+ by pyrophosphatases (PPases) through biomembranes. Membrane PPases couple the hydrolysis of inorganic pyrophosphate to the active transport of cations across membranes. The PPases always require Mg2+ for function (Baykov et al., 1993; Luoto et al., 2011). The hyperthermophilic bacterium Thermotoga maritima, found in hydrothermal environments, as well as the mesophile Methanosarcina mazei contain membrane-bound Na+-pyrophosphatases (Tm-PPase and Mm-PPase, respectively) that are homologs to H+-PPases (Belogurov et al., 2005; Malinen et al., 2008). Both Tm-PPase and Mm-PPase have an absolute requirement for Na+ as well as Mg2+, but display maximal activity in the presence of millimolar levels of K+. Holm & Baltscheffsky (2011) have, therefore, proposed that PPi preceded ATP and Na+ transport preceded the H+ transport in connection with the origin and initial evolution of life on Earth (Fig. 2).
It has been known for some time that hydrogen containing magnesium phosphates like whitlockite (Ca18Mg2H2(PO4)14 and newberyite (MgHPO4·3H2O), that is, protonated orthophosphates, at heating react to form condensed phosphates and water (Sales et al., 1993; Gedulin & Arrhenius, 1994; Arrhenius et al., 1997). The phosphate condensation is attributed to the protonation of the phosphate, because at heating, the hydrogen reacts with one of the oxygen ligands of the phosphorus and leaves as water. As a response, the structure of the orthophosphate () rearranges to form one or more anhydride P–O–P bonds (Arrhenius et al., 1997), that is, the backbone of condensed phosphates. Experiments with newberyite have revealed that dehydration of this crystalline mineral on heating results in an amorphous, more soluble, phase (Sales et al., 1993). During the initial stages of the dehydration process (100–300 °C), condensation to pyrophosphate occurred. At about 200 °C, triphosphate (PPPi) started to appear. However, at 300 °C, a decrease in the concentration of pyrophosphate began, whereas the concentration of triphosphate continued to increase to at least 400 °C, that is, it reaches its maximum in the temperature interval in which dehydration of a subducting slab in a convergent plate margin occurs, that is, 300–375 °C (Alt & Shanks, 2006; Holm & Baltscheffsky, 2011).
A related way of PPi formation can be found in the oxidation of the phosphide mineral schreibersite (Fe,Ni)3P. This mineral is normally referred to as a component of iron meteorites (Pasek, 2008), but it is also known to occur in terrestrial basalts (Pauly, 1969; Ulff-Møller, 1985), as well as an indigenous mineral in lunar basalts (El Goresy et al., 1971). The schreibersite appears to be formed as a by-product to phosphoran olivine in P-rich basalt melts at fast quenching (Boesenberg & Hewins, 2010), and it is possible that the occurrence of this compound is the solution to the ‘phosphate problem’ as discussed by Schwartz (1971, 2006) and Rauchfuss (2008), that is, dissolution of phosphate compounds is necessary before activation can occur. Schreibersite oxidizes slowly in contact with fluid water as the surrounding mineral matrix gets weathered, and forms several phosphorus species of mixed oxidation states like orthophosphate, pyrophosphate, hypophosphate, phospite, etc. (Pasek & Lauretta, 2005; Pasek et al., 2007, 2008; Pasek, 2008). In systems containing dissolved Mg2+ and Ca2+ chloride salts whitlockite is also formed (Pasek & Lauretta, 2005).
Trimetaphosphate () has been shown to be the most effective condensing agent among polyphosphates, particularly because it can serve as a phosphorylating reagent in aqueous and strongly alkaline solution (pH 12) (Etaix & Orgel, 1978; Yamagata et al., 1995; Ozawa et al., 2004). The prebiotic synthesis proceeds by the thermolysis of phosphate to yield linear polyphosphates, which in turn cyclize to trimetaphosphate in aqueous systems (Kalliney, 1972; Ferris & Usher, 1988). Twenty years ago, Yamagata et al. (1991) published a paper on the occurrence of condensed phosphates in fumaroles of the Mount Usu volcano in Japan. The phosphates were believed to have been formed through partial hydrolysis of P4O10. Approximately equal concentrations of pyrophosphate (0.45 μm) and triphosphate (0.37 μm) were found. It is not known whether this was cyclic or linear triphosphate, but in the case of linear triphosphate, it would have cyclized to trimetaphosphate in water.
Decyclization of trimetaphosphate into linear triphosphate, on the other hand, requires the addition of Mg2+ (Turian & Rivara-Minten, 2001). Mechanistically, this can be explained by the shielding of 2+1 PO− groups by two Mg2+, keeping one positive charge free to ionize H2O to H+ and OH−, which would allow nucleophilic attack of one of the three P–O–P anhydride bonds by the OH− (see section ‘Magnesium, nucleotides, and folding’ below). Linear triphosphate is an extremely important constituent of nucleotides in the sense that RNA chains always start with guanosine triphosphate (pppG) or adenosine triphosphate (pppA).
Magnesium, Nucleotides, and Folding
Magnesium plays a special role in the function and folding of nucleotides (Freisinger & Sigel, 2007). As the coordination geometry of Mg2+ is octahedral, it may coordinate six oxygen atoms of different oxyanions. The divalent Mg ion has an enhanced ability compared to other cations to form bidentate clamps with RNA, because the oxyanions of RNA phosphates have significant affinity for Mg2+ (Fig. 3) (Petrov et al., 2011). Unlike other cations, Mg2+ may bring oxygens from two different (charged) phosphate groups in its first shell into direct contact with each other (Hsiao & Williams, 2009). The coordination capacity of Mg2+ is attributed to its size, which makes it possible to simultaneously coordinate negatively charged oxygen of two phosphate groups (Yamagata et al., 1995). Divalent magnesium therefore normally coordinates two adjacent oxygen atoms of pyro- and triphosphate. This is also the reason why, during RNA-folding, Mg2+ can offset the phosphate–phosphate repulsion by being in a central position between the two groups. For the same reason, Mg2+ has been shown to be an extremely effective catalyst for the synthesis of 2′,3′-cyclic AMP and ATP in aqueous solution (Yamagata et al., 1995). Divalent magnesium was shown to accelerate this reaction approximately one hundred times more compared with systems with other cations, like Ca2+ (ibid.). The difference in this respect between Mg2+ and Ca2+ is attributed to the difference in ionic radius. The smaller magnesium ion (0.65 Å, coordination number CN = 6) can form a complex with two or three oxygen atoms of different phosphate groups in a phosphate ester more strongly than a divalent calcium ion (0.99 Å, CN=6) (Yamagata et al., 1995).
Ozawa et al. (2004) have shown that phosphorylation of AMP into ADP and ATP can outrun their hydrolysis in a hydrothermal environment (100–130 °C) if trimetaphosphate is used as the phosphate source. The degree of phosphorylation showed a strong positive correlation to an increase in Mg2+ activity. The recovery of ATP was higher than ADP at 100 °C, while the opposite was true at 130 °C.
Another effect of magnesium’s ability to coordinate oxygen in phosphate groups is that orthophosphate is a strong inhibitor of brucite dissolution at pH>10.5 (Pokrovsky et al., 2005). Brucite may quantitatively co-precipitate orthophosphate, ATP, and most other phosphorylated organic compounds (Karl & Tien, 1992). In DLH, which derive from brucite, anions of interest in the context of molecular evolution have a sequence of binding energies Cl− < < < < linear oligophosphates (like PPPi)<cyclic oligophosphates (like trimetaphosphate) (Arrhenius et al., 1997). Choy et al. (1999) have demonstrated that nucleotides such as CMP, AMP, GMP, and even DNA can hybridize with DLH, giving rise to a heterostructured nanohybrid which consists of one nucleotide layer and Mg(OH)2 alternatively. The nucleotides stabilized in the interlayer space of DLH retain their chemical and biological integrity. Thermal stability of the encapsulated nucleotides up to 300 °C has been observed (Choy et al., 2004).
The Chemistry of Nucleotides May Include Borate as an Active Compound
Saladino et al. (2011) have recently shown that formamide, the hydrolysis product of HCN, oligomerizes in the presence of borate minerals yielding, at the same time, nucleobases and carboxylic acids. Their experiments were carried out at 160 °C in the presence of about 20 different borate minerals and thus yielded the building blocks of both the genetic and metabolic apparatuses. Particularly, the Na-Mg-Al borate mineral dravite (NaMg3Al6(BO3)3Si6O18(OH)4) was shown to be efficient in promoting the formation of purine (19% recovery). The efficiency in purine formation in the presence of dravite was probably due to the oligomerization of HCN to its tetramer diaminomaeleonitrile (DAMN), which was shown to be present in the experimental system. DAMN is an intermediate in the abiotic formation of purines, pyrimidines, as well as amino acids from HCN (Ferris & Usher, 1988). Hydrogen cyanide, in turn, may be formed from ammonia and methane or carbon monoxide in oceanic basement and then be concentrated on secondary mineral surfaces or as ferrocyanide (Holm & Neubeck, 2009).
Boron has never been in focus of geobiology because it is not a major component of biological macromolecules. Boron has, however, strong affinity for organic material because it forms trigonal and tetrahedral complexes with oxygen groups. Borate readily forms complexes with a wide variety of sugars and other compounds containing cis-hydroxyl groups, such as humic substances (Bolaños et al., 2004; Panagiotopoulos & Sempéré, 2005). Seawater is the source of almost all borate in altered oceanic crust. In this sense, it is similar to phosphorus.
Prebiotic Synthesis and Stability Conditions of Nucleotide Constituents
LaRowe & Regnier (2008) have calculated the thermo-dynamic potential for abiotic synthesis of the five common nucleobases (adenine, cytocine, guanine, thymine, and uracil) and the two monosaccharides (ribose and deoxyribose) in RNA and DNA from the precursor molecules CH2O and HCN. The temperature, pressure, and bulk composition conditions were chosen to be representative for hydrothermal conditions in ultramafic settings like the Lost City hydrothermal system (Delacour et al., 2008). All of these biomolecules were shown to be thermodynamically favored to be synthesized throughout the temperature range from 0 °C to between 150 and 250 °C, depending on the single biomolecule.
Mechanistically, Yuasa & Ishigami (1977) have shown that periclase (MgO) and magnesite (MgCO3) serve as deprotonating reagents for the condensation of HCN to DAMN. At least systems with MgO proceed with further reactions to amino acids and heterocyclic nitrogen compounds like adenine and 4-aminoimidazole-5-carboxyamide (AICAI). In this context, it is notable that Ferris and co-workers (see, for instance, Ferris, 1984) have worked out a prebiotic synthesis pathway from DAMN via AICAI to a wide array of purines, including adenine and guanine.
Pentoses, like ribose, can be formed by the formose reaction under strongly alkaline conditions from simple organic precursors (formaldehyde and glycolaldehyde) (Ferris, 2005). For a while, the formose reaction was an outdated concept in prebiotic chemistry because the reaction was considered to be non-selective. However, it is now believed by many scientists that a selective mechanism that favors the formation of ribose has been identified (see below). The formation of monosaccharides proceeds by the stepwise condensation of formaldehyde to a dimer (glycolaldehyde), trimer, etc. Aldehydes can be formed directly from elemental carbon in the presence of water (Flanagan et al., 1992). Elemental carbon in the form of graphite and amorphous carbon is common in peridotites (Delacour et al., 2008). The initial reaction of elemental carbon with water gives hydroxymethylene, which can rearrange to formaldehyde. A new hydroxymethylene molecule can then add onto the formaldehyde (as well as larger aldehyde molecules) and form glycolaldehyde.
Delidovich et al. (2009) have carried out condensation reactions of monosaccharides with formaldehyde, glycolaldehyde, and glyceraldehyde in, what they call, ‘a weakly alkaline medium’ (pH 10.4) at 70 °C in the presence of MgO as a catalyst. The condensation of glycolaldehyde and formaldehyde in the presence of the hydrated MgO at pH 10.4 resulted in the formation of trioses, tetroses, and pentoses.
It has been shown that borate minerals stabilize ribose and simultaneously favor its five-membered furanose ring (Prieur, 2001; Ricardo et al., 2004; Amaral et al., 2008). As previously mentioned, borate readily forms complexes with a wide variety of sugars and other compounds containing cis-hydroxyl groups (Bolaños et al., 2004; Panagiotopoulos & Sempéré, 2005). Once formed, the cyclic form of the pentose like ribose forms a stable, less reactive complex with borate. The binding preferences of borate to pentoses has been determined to be ribose>lyxose>arabinose>xylose (Li et al., 2005). NMR analysis shows that borate occupies the 2 and 3 positions on the furanose form of ribose, thus leaving the five position available for potential reactions like phosphorylation (Benner, 2004; Ricardo et al., 2004). Borate-bound ribose is formed preferentially at pH>9, because the predominant boron species at lower pH is boric acid (Amaral et al., 2008). It has also been shown to be stable for long periods of time at temperatures of about 60 °C (ibid.). In biological systems, the purine nucleotides are synthesized by constructing the purine base on a pre-existing ribose-5-phosphate (Joyce, 1989). However, results by Etaix & Orgel (1978) show that adenosine-5′-triphosphate can be synthesized directly from adenosine and trimetaphosphate at pH 12 if the 2 and 3 OH groups of the ribose are blocked by borate.
Prieur (2009) has shown that it is possible to phosphorylate pure ribose with trimetaphosphate to ribose-5-phosphate in the presence of borate salts. According to Prieur, pyrophosphate does not appear to have the same phosphorylating effect. At high pH (pH>10), the nitrogen in the N9 position of adenine loses a proton so that adenine becomes an anion (Gonella et al., 1983). The adenine anion may react with the C1 position of a pentose-5-phosphate in the open form with the aldehyde group exposed to attack by a nucleofile and form a β-glycosidic bond (Mellersh & Smith, 2010). In order then to stabilize the product, a furanose or pyranose ring needs to be formed, so only pentoses or longer carbohydrates could participate in this type of reactions. This pH effect may support a notion that prebiotic purine nucleotides were, in principle, synthesized in the same way like in modern biological systems by constructing the purine base on a pre-existing ribose-5-phosphate (Holm et al., 2010).
Ribozymes and Ribosomes
Ribozymes are ribonucleic acid enzymes that are capable of both catalyzing chemical reactions and storing genetic information (Cech et al., 1981; Gilbert, 1986). These two features support the hypothesis of a pre-protein RNA world (Gilbert, 1986; Bereźniak et al., 2010). In the first report on catalytic RNA (or a ribozyme) by Thomas Cech’s group, incubation of the pre-rRNA had been carried out in the presence of Mg2+ (Kruger et al., 1982; Cech, 2002). A principal feature of RNA is its high negative charge density. Therefore, the presence of positively charged ions is essential for the structural integrity and catalytic activity of all ribozymes, as well as for their cleavage (Sreedhara & Cowan, 2002; Walter & Engelke, 2002; Lindh, 2005). In the cell, the favorite cation of choice is Mg2+, as it has such high affinity for the negatively charged backbone of RNA and is the most abundant divalent metal ion.
The ribosome, which synthesizes proteins in cells, is one of the most ancient molecular machines in all living systems. It is believed to have achieved its present form before the emergence of the last common ancestor of modern life (Fox & Naik, 2004). The ribosome is actually a large ribozyme in which the catalytic center is still composed of RNA (Walter & Engelke, 2002). It has been known for decades that Mg2+ is involved in both the structure and catalytic activity of ribosomes (Klein et al., 2004; Lindh, 2005). The core region of the ribosomal large subunits (LSUs), defined with the site of peptidyl transfer (PT) as the origin, is characterized by the greatest Mg2+ density and the lowest ribosomal protein density (Hsiao et al., 2009). As one moves from the center to the periphery of the LSU, proteins replace magnesium ions. Near the PT-origin, phosphate oxygens more frequently act as inner sphere Mg2+ ligands (Hsiao et al., 2009). It is, therefore, tempting to interpret the density distribution of Mg(II) as a record of chemical evolution of this ancient cellular machinery. Divalent magnesium would have been responsible for the functional folding of early RNA and has successively over time been replaced for the same purpose by the more sophisticated proteins.
High pH in Plate Tectonic Convergent Margins
Organic precursor molecules such as NH3, CO, and HCN are strongly adsorbed on secondary minerals like smectites and zeolites during seawater circulation at the slightly acidic pH of basalts of oceanic crust (Fripiat et al., 1972; Gualtieri, 2000; Colín-García et al., 2010). Once adsorbed, these molecules will be carried passively for perhaps millions of years along with the oceanic plate toward a subduction zone, as illustrated by the adsorption of methane, ethane, and propane to deep sea clays and their desorption by the addition of strong base (Hinrichs et al., 2006). Convergent margins in the form of plate tectonic subduction zones are the most dynamic regions on Earth. The Mariana forearc in the western Pacific Ocean, where the Pacific plate subducts beneath the Philippines plate, is a non-accretionary forearc with numerous serpentine-brucite and carbonate mud volcanoes next to a deep ocean trench (Fig. 4). Near the Mariana trench, that is, at a lateral distance of 48–54 km from the maximum depth of the trench into the overriding Philippines plate, the upwelling pore waters of the Mariana forearc have pH of 10.7 and are less saline than the ambient seawater, because the waters originate by dehydration of the subducting Pacific slab at temperatures of 300–375 °C (Alt & Shanks, 2006; Mottl, 2009). These proximal springs form chimneys of the mineral brucite on the seafloor. Farther from the trench (70–90 km lateral distance), the fluid chemistry changes abruptly and the waters have pH 12.5 and are more concentrated with respect to dissolved inorganic species relative to seawater (Mottl, 2009). These distal springs form chimneys of aragonite and calcite. Serpentines are thus permeated by carbonate-rich, high-pH hydrothermal solutions at medium temperature (100–300 °C). The reason that the fluids close to the trench have a pH of about 10.7 is because of the consumption of H+ during serpentinization and brucite formation of primary silicate minerals (Holm & Neubeck, 2009). Mg(OH)2 is, in fact, excellent at buffering pH at alkaline conditions and has been used for that purpose in prebiotic peptide synthesis experiments (Huber et al., 2003). However, the pH of 12.5 of the distal pore fluids requires an additional explanation, which is still under debate.
The fossil record
It is believed that biological processes started to affect Earth’s surface mineralogy by 3.85–3.6 Ga (Hazen et al., 2008). The Archean sedimentary record indicates that areas of shallow water on continental crust existed already by 3.6 Ga (Nisbet, 1987), and serpentine mud volcanoes similar to those of the Mariana forearc occur at Isua, Greenland, and have been dated to early Archean (3.81–3.70 Ga) (Pons et al., 2011). Parts of the shallow seas were floored by evaporites, which show that the early ocean was not fresh: Its composition was in part controlled by hydrothermal circulation through spreading-centers consisting of komatiitic lavas (>18% MgO) (Nisbet, 1987; Schoonen et al., 2004). Magnesite deposits that occur in Archaean strata are of controversial origin, although Nisbet (1987) suggests that they may be a variety of evaporite precipitated from fluids rich in magnesium and carbonate. However, the magnesites that exist in Isua are suggested to be precipitates from carbonate-rich fluids in subduction zones (Dymek et al., 1988; Pons et al., 2011).
Another group of evaporitic minerals are borates, which may be either marine or geothermal (‘non-marine’). Borate is rapidly taken up from seawater during low-temperature alteration of the oceanic crust (Palmer, 1991; Palmer & Swihart, 1996). It is enriched in serpentinites and basalts altered by seawater at relatively low temperatures (Bonatti et al., 1984; Seyfried et al., 1984). At 150 °C and below, boron is removed from seawater and is incorporated into alteration mineral phases like celadonite and Mg-smectite (Alt, 1995). A complete balance of the ocean boron budget requires deeply subducted boron to be returned efficiently to the oceans via arc volcanism. Several lines of evidence indicate that most of the boron is recycled from the subducted slab at convergent margins (You et al., 1993), like the Marianas and the Lesser Antilles arcs (Mottl, 1992; Stout et al., 2009).
Grew et al. (2011) reviewed the history and evolution of boron minerals to evaluate whether borate minerals could have been present to stabilize pentose sugars on the early Earth. The minerals that were primarily discussed in their paper were chosen on the basis of the experiments by Ricardo et al. (2004), in which ribose was stabilized in the presence of the marine evaporitic borate minerals ulexite (NaCaB5O9·8H2O), kernite (Na2B4O7), or colemanite (Ca2B6O11·5H2O). Grew et al. (2011) concluded that although there is no firm evidence that any of these studied minerals was present prior to 3.7 Ga, there could have been sufficient B for stabilizing ribose in the ocean if plate tectonics were active. Because the geological evidences known today support an initiation of plate tectonics on Earth in the Hadean eon, or before about 4 Ga ago (Harrison, 2009), a scenario with deeply subducted boron returned efficiently toward the seafloor via subduction of oceanic slabs and arc volcanism is highly likely. The oldest reported colemanite and ulexite deposits are only 330 Ma, and the oldest reported kernite just 19 Ma. However, 2.4–2.1 Ga borate deposits with an evaporitic origin exist in the Liaoning and Jilin Provinces, China. The borate mineralogy here is dominated by the magnesium borate minerals szaibelyite (Mg2B2O5·H2O), suanite (Mg2B2O3), with minor ludwigite ((Fe,Mg)4Fe2B2O7) associated with a sequence of metasediments and metavolcanics (Peng & Palmer, 2002). Suanite and szaibelyite are believed to have formed from dehydration of original hydrated borates like inderite (Mg2B6O11·15H2O). Geological, geochemical, and boron isotope studies indicate that these borates are metamorphosed evaporites of geothermal brines rich in B and Mg. The tectonic evolution of this area started with subduction of oceanic crust below an Archean craton, continental collision, and development of extensional basins. Geothermal fluids associated with the tectonic activity leached the oceanic crust that was in the process of subduction for enriched components, like boron and magnesium, and were channeled along fractures and faults through the overriding plate, in analogy to the processes in the Mariana and Lesser Antilles arc systems.
Modern convergent margins
Serpentine-brucite type mud volcanism in convergent margin settings is not merely a local curiosity of the Mariana system but occurs worldwide and have done so through Earth’s history (Salisbury et al., 2002; Hyndman & Peacock, 2003; Pons et al., 2011). The forearc mud volcanoes are fed by a rising column of serpentinized peridotite that is mobilized by water and volatiles from the subducting plate. The mud volcanoes produced in this way accumulate large amounts of secondary minerals and can be big – about 2 km high and 50 km across. The Mariana forearc is of particular interest in the context of prebiotic organic synthesis because it is non-accretionary with respect to ocean floor sediments and the interaction between water and mafic rocks is not obscured by other processes. The boron concentration of pore water from ODP Site 1200 on the summit of the South Chamorro Seamount in the Mariana convergent margin has been shown to be 4 mmol kg−1 compared to 0.4 mmol kg−1 in seawater (Mottl et al., 2003). Notably, the highest reported B concentration in geothermal springs of the St. Lucia island of the Lesser Antilles convergent margin is 135 mmol kg−1 (Stout et al., 2009). Therefore, these kind of systems are probably the best modern analogs to the fossil environments in Liaoning and Jilin. The sedimentation rate of pelagic material is slow and ‘contamination’ by both biogenic and minerogenic particles is at a minimum. Interstitial fluids of pH 12.5 associated with serpentinized mud are enriched in dissolved components like carbonate, light hydrocarbons, borate and ammonia (Mottl et al., 2003).
Phosphates are scavenged from seawater by ridge as well as ridge-flank hydrothermal activity and are accumulated in oceanic lithosphere in analogy to borates. Wheat et al. (1996) have estimated that ridge-axis and ridge-flank hydrothermal processes in combination remove about 50% of the global input of dissolved phosphorus from rivers into oceanic lithosphere. The occurrence of phosphate minerals is generally considered rare in ultramafic rocks. However, Chopin et al. (2001) report on the occurrence of raadeite (Mg7(PO4)2(OH)8) and other Mg-phosphates in nodules of apatite of the Tingelstadtjern serpentinite body, Norway.
In subduction zones, distillation, compaction, and smectite-to-illite transition are likely to be responsible for the liberation of organic precursor molecules. Because the fluids also carry the particulate matter that becomes the ‘serpentinite mud’ volcanoes, different types of active mineral surfaces are carried along. Analyses of the pore fluid composition show that the ultramafic rocks of the overriding Philippines plate in the Mariana system have reacted with a fluid derived from dehydration of the subducted slab of the Pacific plate, and not by seawater (Mottl, 1992). Subduction of oceanic lithosphere cools the overlying forearc close to the trench, so that the forearc mantle rocks get exceptionally cool, that is, ∼100 °C (Hyndman & Peacock, 2003). The alkaline pH of these subduction environments releases molecules like HCN and allows prebiotic synthesis of amino acids, carbohydrates, and heterocyclic nitrogen bases, etc. (Holm & Neubeck, 2009). The occurrence of cyanides in hydrothermal fluids of convergent margins has so far been reported from the Kurile Islands and the Kamchatka Peninsula (Mukhin, 1974; Mukhin et al., 1978). The synthesis of these prebiotic key compounds and their stability are promoted by the prevalent moderate temperatures, as shown by the experiments by Ozawa et al. (2004) and LaRowe & Regnier (2008) (100–130 °C and 0–250 °C, respectively). Released fluid carrying HCN and formamide can be expected to rise from an environment of relatively low pH into hydrated mantle of high pH. At the high pH, HCN monomers may form oligomers. HCN and formamide may form amino acids as well as purine and pyrimidine bases. Formamide may also form carbohydrates, which may merge with nitrogen bases and form nucleosides. These may, in turn, be phosphorylated to nucleotides by pyrophosphate or trimetaphosphate present in the system. The nucleotides – the constituents of RNA – may eventually hybridize with the secondary DLH and in this way be stabilized for further prebiotic reactions in the interlayer space of the DLH.
John Maddox writes in a note in his book ‘What Remains to be Discovered’: ‘The close similarity of the chemical structure of the RNA molecules of the ribosomes from organisms of the same class is persuasive evidence of the antiquity of ribosomes in the evolution of living things’ (Maddox, 1999). Obviously, the ribosome is one of life’s most ancient machines. Magnesium is an essential element for any living organism today and has, no doubt, been so through life’s entire history on Earth. The fact that the biogeochemistry of magnesium is intimately coupled to that of phosphorus and nucleotides via the chemistries of boron and carbohydrates, indicates that it also had a key position in the prebiotic processes leading to the origin of life. Anyone who favors the notion that a pre-RNA world has existed must take its requirement for divalent magnesium of ribozymes and early enzymes like PPases for function into consideration. The major reason for its importance is that the properties of divalent magnesium are generally different from those of all other cations. Divalent magnesium has the largest hydrated radius of any common cation, but at the same time the ionic radius is among the smallest. Magnesium has always been abundant on Earth in the form of both monovalent (Mg(OH)+ and divalent cations (Mg2+) as well as hydrated complexes and secondary mineral phases in weathered seafloor. Borates and phosphates are accumulated in the same type of settings. Pyrophosphate and trimetaphosphate that can be active in prebiotic processes may be formed in environments of high pH and low water activity, like in forearc mantle rocks. Taken together, these circumstances point to the oceanic lithosphere as a likely environment for the initiation and propagation of life processes, particularly those parts of the ocean floor that were in the early phase of subduction at convergent margins.
This work has been supported by the Stockholm University Astrobiology Centre (SU-ABC) and the Stockholm University Faculty of Science. The author has no conflict of interests.