sign in sign up
<<
Home , Sandra Pizzarello, Tholin

Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

Planetary Organic Chemistry and the Origins of

Steven A. Benner, Hyo-Joong Kim, Myung-Jung Kim, and Alonso Ricardo

Foundation for Applied Molecular Evolution and The Westheimer Institute for Science and Technology, Gainesville, Florida 32601 Correspondence: [email protected]

Organic chemistry on a planetary scale is likely to have transformed and reduced carbon species delivered to an accreting . According to various models for the origin of on Earth, biological that jump-started Darwinian evolution arose via this planetary chemistry. The grandest of these models assumes that ribonucleic acid (RNA) arose prebiotically, together with components for compartments that held it and a primitive metabolism that nourished it. Unfortunately, it has been challenging to iden- tify possible prebiotic chemistry that might have created RNA. Organic molecules, given energy, have a well-known propensity to form multiple products, sometimes referred to col- lectively as “tar” or “tholin.” These mixtures appear to be unsuited to support Darwinian processes, and certainly have never been observed to spontaneously yield a homochiral genetic polymer. To date, proposed solutions to this challenge either involve too much direct human intervention to satisfy many in the community, or generate molecules that are unreactive “dead ends” under standard conditions of temperature and pressure. , organic species having carbon, , and atoms in a ratio of 1:2:1 and an aldehyde or ketone group, conspicuously embody this challenge. Theyare com- ponents of RNA and their reactivity can support both interesting spontaneous chemistry as part of a “ world,” but they also easily form mixtures, polymers and tars. We describe here the latest thoughts on how on this challenge, focusing on how it might be resolved using minerals containing borate, silicate, and molybdate, inter alia.

nteresting organic chemistry occurs through- certainly contributed to the reduced carbon Iout the cosmos, including in presolar nebulae inventory on Earth before life emerged, plane- (see the article in this collection by Pascale Erh- tary processing on Earth undoubtedly also con- renfreund), asteroidal bodies (see the article in tributed to the inventory of prebiotic molecules this collection by Sandra Pizzarello) and icy that were available to life as it originated (as- bodies near the outer boundary of our solar sys- suming that Earth was the site of life’s origin). tem (Bernstein et al. 2002). Although organic Indeed, in the RNA first model for the origin molecules made in off-Earth locales almost of life on Earth (Joyce and Orgel 1999)(Benner

Editors: David Deamer and Jack W. Szostak Additional Perspectives on The Origins of Life available at www.cshperspectives.org Copyright # 2010 Cold Spring Harbor Laboratory Press; all rights reserved; doi: 10.1101/cshperspect.a003467 Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467

1 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

S.A. Benner et al.

2009), it is often proposed that terran-based organic transformations might have occurred on chemistry produced RNA in oligomeric form dry land or below on a to initiate Darwinian evolution. that was totally submerged. How are we to constrain models for planetary processing to converge on a model for what actually happened on Earth four billion years BACKGROUND ago? Today, atmospheric dioxygen (O2)readily Organic Molecules with Energy converts organic materials to carbon dioxide, Spontaneously Yield Polymers and Complex making it essentially impossible to observe such Mixtures processing on the surface of Earth. Furthermore, the ubiquity of life on modern Earth means that Any model for planetary organic chemistry must any organic processing is more likely to reflect recognize that very few organic molecules are biology than prebiology. The closest we may thermodynamically stable in water, either with come today to observe organic transformations respect to conversion to their fully hydrated state absent biology on a planetary scale might be on or decomposition upon heating to elemental car- , a moon of whose atmosphere and bon (charcoal), carbon dioxide, orother thermo- surface is rich in reduced carbon. dynamic “end points.” Accordingly, any model Nevertheless, it is possible to apply a general for planetary prebiotic chemistry must address understanding of organic chemical reactivity to the metastability of organic species. This word suggest chemical reactions that might have captures the concept that organic molecules that occurred on early Earth and the products that have appeared through the interaction of precur- they might have produced. These suggestions sors with energy, water, and other organics can are constrained by models for the atmosphere then disappear upon further interaction. As and mineralogy of early Earth, although these many authors have noted (Cairns-Smith 1982; constraints might change as models improve. Shapiro 1987; Shapiro 2007), any prebiotic reac- In this article, we assume that the atmos- tion scheme that requires two or more organic phere of early Earth was less oxidizing than species must be concerned about the metastabil- today’s atmosphere, although not as rich in ity of two or more species; any scheme that does as the simulated atmosphere used in not produce both components in useful concen- the classic experiments of Stanley Miller (Miller trations at the same time will not meet a standard 1955). Further, we assume that the atmosphere of proof that the community need to accept a on early Earth had access to many sources solution to the problem. of energy. These include electrical discharge, Here, too, the mineral inventory of early and visible light (although the Sun Earthcannotbeignored.Mineralsofmanykinds was almost certainly dimmer then than now, a may have guided the reactivityof organic species Titan-like haze may have prevented high energy that emerged on early Earth, altered their meta- photons from reacting the Earth’s surface), stability, and influenced the time when specific volcanism (providing not only heat but also organic species were available to emerging life. reactive species and mixtures not at thermo- dynamic equilibrium), ionizing radiation, and Carbohydrates Embody this Natural impacts.(SeePizzarelloandShock2010foradis- Propensity to Polymerize cussion of such energy sources.) We also assume that life emerged after Toexplorethese points anddevelop the scientific the planet underwent a geological fractionation methods that enable this exploration, we will use in which heavier minerals and elemental iron carbohydrates as a focus. Carbohydrates are sank towards the core, leaving lighter rocks to organic species having carbon, hydrogen and formthecrust.Openquestionsconcerntheinven- oxygen atoms in a ratio of 1:2:1. They are there- tory of water relative to the surface of early Earth, fore at the same oxidation level of elemental car- an inventory that determined whether planetary bon. Furthermore the simplest carbohydrate,

2 Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

Planetary Organic Chemistry and the Origins of Biomolecules

(HCHO, H2C¼O, or C1H2O1)is and preclude certain compound classes based easily generated by electrical discharge or ultra- on the same data, characterizes the “origins” field violet radiation impinging on moist atmos- in general (Benner 2009). To evaluate such con- pheres that are rich in carbon dioxide; nearly tradicting views, participants in this field must every contemporary model for early Earth per- understand relevant features of bonding and mitssuchatmospheres(Pintoetal.1980;Cleaves reactivity in organic molecules. This understand- 2008).Becausetheycontainketoneandaldehyde ing requires in turn that they understand the for- groups (see later), carbohydrates have interest- malisms used by organic chemists to describe ing reactivity that includes the ability to form bonding and reactivity in organic molecules. newcarbon-carbonbondsunder “standardcon- Although the features and formalisms are taught ditions,” defined as those where water is liquid in introductory organic chemistry courses, it is under contemporary terran atmospheric pres- worth summarizing them here a prelude to dis- sure. By comparison, the compounds that often cussing constraints on reactions that might be concern prebiotic chemists (carboxylic acids, expected to occur in prebiotic environments. fatty acids, and amino acids, for example) have First, a single bond joining two carbon essentially no such reactivity. Indeed, carbohy- atoms is strong, on the order of 400 kJ (100 drates are “high energy” because they can rear- kcal) per mole. This means that a pair of carbon range their constituent atoms to give carboxylic atoms joined by a typical single bond will acid derivatives and other more stable end point remain joined for many millions of years at tem- “sinks.” peratures when water is a liquid at sea level on Recognizing this, some authors, most nota- Earth near neutral pH (“standard conditions”). bly Arthur Weber (Weber 2001a; Weber 2001b; This is also true for single bonds between car- Weber 2007) (see Blackmond 2010), have ex- bon and hydrogen. ploited the reactivity, energy, and prebiotic In contrast, single bonds between carbon accessibility of carbohydrates to suggest entire and oxygen carbon and , although sim- “carbohydrate world” metabolisms at or near ilarly strong, but tend to be less persistent over life’s origins. This exploits the energy of formal- time because they confer reactivity upon dehyde, hydroxyaldehydes and hydroxyketones organic species that possess them. Heteroatoms relative to isomeric forms that have carboxylate (meaning neither carbon nor hydrogen) create groups, to do chemistry, much of it reminiscent centers of reactivity that are “weak” spots in of modern metabolism. organic compounds. In particular, they provide Unfortunately, others use the very same paths under standard conditions where bonds reactivity and energy to argue that carbohy- between carbon atoms or between carbon and drates could not have been present on early hydrogen atoms can be broken at the same Earth (Shapiro 1988). For example, based on time as a bond to another atom is formed. the short survival time of ribose, especially at This compensates the energy lost in the break- high temperature and high pH, Stanley Miller ing bond with energy gained with a simultane- and his coworkers (Larralde et al. 1995) con- ously forming bond, allowing the reaction to cluded that “ribose and other sugars were not occur under relatively mild conditions. components of the first genetic material” and precluded their presence in prebiotic scenarios. Pairs of Electrons form Bonds between Atoms This is despite the fact that simple carbohy- drates such as are well known Understanding ways that new bonds can form in the cosmos (Hollis et al. 2000). as old bonds are breaking is a key to understand- ing what reactions actually will occur under standard conditions. Recalling general chemis- ORGANIC CHEMISTRY SKILLS: A REMINDER try, covalent bonds between two atoms are This apparent paradox, where experts within formed by the sharing of pairs of electrons the same community simultaneously exploit (Fig. 1). Thus, in water (H-O-H), the lines

Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 3 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

S.A. Benner et al.

HH HH = HH React Two hydrogen To for m a dihydrogen atoms, each with one electron with the two hydrogen nuclei held together by two electrons

O O = H OH = H O H HHC HHC

Water molecules are formed In formaldehyde the bond between by the reaction of two hydrogen the carbon and oxygen is held by nuclei with an oxygen nuclei 4 electrons.

Figure 1. Lewis structures of chemical bonds for dihydrogen (H2), water (H2O) and formaldehyde (H2C¼O, or CH2O, or HCHO).

between the hydrogen atoms and the central A Nucleophilic Center Brings a Pair of oxygen atom each represent a pair of electrons Electrons to Form a New Bond that form a single bond holding the two hydro- As a chemical bond is a pair of electrons gen atoms to the oxygen. In formaldehyde between two atoms, any unshared pair of elec- (H C¼O), the double line between the carbon 2 trons is available in principle to form a new and the oxygen represents two pairs of elec- bond. Further, if a bond breaks, the electrons trons; four electrons in total bind the C atom that were in that bond are available to form a to the O atom. new bond. Atoms that contain pairs of electrons Chemists frequently do not write symbols available to form a new bond (or can get them representing all of the atoms and electrons. by breaking a bond) are called nucleophilic Therefore, the first skill required in analyzing centers (Fig. 2). reactivity in organic chemistry requires that To form a bond, the electron pair from the one put back into a structure the symbols that nucleophilic center must find an atom that practiced chemists do not write (but under- has a vacant orbital or can get one by losing a stand are there), completing the structure of bond through breakage. This atom is called an the organic molecule. This ensures that the electrophilic center. An archetypal electrophile analysis reflects all of the atoms and electrons þ is a proton (H ). A proton is not bonded to in a molecule. anything, has a vacant 1s bonding orbital, so it Astructurecompletedinthiswayisknownas can form a bond with a single partner (Fig. 3). a Lewis structure. A Lewis structure explicitly An archetypal nucleophilic center is the oxygen indicates electrons that are not involved in bond- atom on water. It carries an unshared pair of ing. For example, in the Lewis structure of water, þ electrons that can form a new bond to H . oxygen carries two pairs of unshared electrons The product of the reaction between water and from the outer valence shell. We represent each þ þ H is H O (the hydronium ion). of these valence electrons not involved in a 3 bond by a dot. Hence, the oxygen in the Curved Arrows Describe the Movement of H-O-H structure has four dots, representing Pairs of Electrons electrons on the oxygen but not involved in bonding. Likewise, the oxygen in formaldehyde Organic chemists use curved arrows to describe carries two pairs of unshared electrons, repre- reactions between nucleophilic and electro- sented again by four dots on the oxygen in the philic centers that produce a new bond. The Lewis structure. curved arrow begins with an unshared pair of

4 Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

Planetary Organic Chemistry and the Origins of Biomolecules

Nucleophilic center O R R H R R C C N Resonance C N N R R R R H H O H Electrophilic center Nucleophilic center

O O R H H H H Resonance C C C R H C C H H C C H

R H H H H

Electrophilic center Electrophilic center Figure 2. Nucleophilic centers have an unshared pair of electrons that can form a new bond, or can get one (via resonance, for example). Electrophilic centers have a vacant orbital (or can get one via resonance, for example) that can accept an unshared pair of electrons from a nucleophilic center to form a new bond. In a resonance form a pair of electrons moves between adjacent atoms (electrons are dynamic entities), creating a new representation for the same molecule. The resulting resonance forms are joined by a double headed arrow to indicate equivalence.

electrons on the nucleophile, the pair that less easy to spot when they get the electron will form the new bond in the product. The pair needed to form a future bond by breaking arrow is drawn to end at a position (on the an existing bond. The same is true for electro- structures of the reactants) where the electron philic centers that obtain vacant orbitals avail- pair will be after the bond is formed. Figure 3 able for forming new bonds only after an shows the reaction of the unshared pair of elec- existing bond is broken. trons on the oxygen of water (the nucleophilic For example, the carbon of formaldehyde þ center) with H (the electrophilic center) to (H2C¼O) has all of its four valences occupied. þ give H3O . That carbon does not seem to have a valence When nucleophilic centers bear an electron available to form a new bond with anything. pair prominently placed in a correctly drawn If, however, one of the two bonds between Lewis structure, they are easy to spot. They are carbon and oxygen breaks, with the electron pair moving from a position between the car- bon and the oxygen to a new position on the oxygen, then the carbon center has a valence Electrophilic H free and can undergo nucleophilic attack by H center + an oxygen atom, for instance, from a nearby + H O molecule. O H O H 2 This process is shown using curved arrows in Figure 4. Here, a bond between carbon and H H Nucleophilic oxygen is broken at the same time as the carbon center forms a new bond to an incoming oxygen atom. At the same time as the energy in the Figure 3. Reaction of the nucleophilic center on the oxygen of water with an electrophilic center, Hþ. second C-O bond is lost through breakage, the The movement of a pair of electrons in the reaction energy of a new C-O bond is gained. The result- is illustrated using a curved arrow. The result is ing product is the hydrate of formaldehyde þ H3O , the hydronium ion. (H4CO2).

Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 5 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

S.A. Benner et al.

– O O Vacant orbital Resonance on this carbon + C C H H H H Formaldehyde

H – O Attack of O Rapid transfer O nucleophile of protons C C C H H H H H H O+ O O H H H H H Hydrate of formaldehyde

Figure 4. Reaction of the nucleophilic center on the oxygen of water with an electrophilic center, the carbon atom of formaldehyde forms the hydrate of formaldehyde. The movement of a pair of electrons in the reaction is illustrated using a curved arrow.

CURVED ARROW MECHANISMS FOR THE After the transfer of some of the hydrogen PREBIOTIC SYNTHESIS OF BIOLOGICAL atoms, the nitrogen of the amino alcohol again MOLECULES has an unshared pair of electrons, and is able to The curved arrow tool can be used to describe form a second bond with the carbon atom. The most reactions of organic molecules under resulting compound is known as an imine (4), ¼ standard conditions. This includes transforma- which contains a C N unit having a carbon tions that might have converted formaldehyde atom bonded twice to a nitrogen atom. ¼ in a prebiotic world into molecules that are The carbon of the C N unit is also an characteristic of contemporary terran life. electrophilic center. This sets the stage for the next reaction, where the carbon of the anion (5) attacks the imine (4) carbon to Curved Arrow Mechanisms for the Formation form an aminonitrile (6). The has a of Amino Acids C;N unit, where the nitrogen is bonded three The curved arrow formalism can be used to de- times to the carbon. This carbon is again an scribe the synthesis from formaldehyde, ammo- electrophilic center. If a pair of electrons nia, cyanide and water of a simple , forming one of the bonds between carbon and þ (NH2CH2COOH), one of the basic nitrogen leaves to form a bond with H , then building blocks of proteins. The steps are shown the carbon has a free valence. It is therefore in Figure 5 with species both named and num- available to form a bond with a nucleophilic bered for future reference. oxygen atom from water. (2) has three hydrogen atoms The product, again after Hþ atoms are trans- and one nitrogen atom. A Lewis structure shows ferred, has another C¼O group in a unit known that the nitrogen in ammonia also carries an as an amide (7). The carbon atom of the amide is unshared pair of electrons. The nitrogen atom againanelectrophiliccenter;itcanbeattackedby is therefore a nucleophilic center. Ammonia the nucleophilic oxygen of another water mole- should therefore react with formaldehyde (1) cule. This leads to the hydrolysis (taking on a for the same reason that water does. In this reac- water molecule) of the amide and the formation tion, the unshared pairof electrons on ammonia of the amino acid glycine, together with an forms a new bond between its nitrogen and the ammonia molecule. carbon of formaldehyde, just as the pair of elec- The net process is the reaction of one mole- trons forming the second carbon-oxygen bond cule of formaldehyde, one molecule of hydro- leaves to form a new bond between the oxygen gen cyanide, and one molecule of water to and Hþ. This generates an “amino alcohol” (3). give one molecule of glycine (8). In terms of

6 Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

Planetary Organic Chemistry and the Origins of Biomolecules

3 O H H H H Formaldehyde 1 O O C H H H C H H C H N H NH H –H N H H Ammonia 2 H H H Aminoalcohol 3 H H H H O C H N N C H C H –H2O H H H H N –H Imine 4 H H

Aminonitrile 6 H H H H H H N C N C –H H N C H H – H H C N H CN C N H H H O cyanide Water O H anion 5 H H Amide 7 H H H H O H –H H N C H N C N C H H C N H H H C N –H C +H H O O N H H H H H H H

H H O H H H H O O H C H H H H N N N C C N N C N H H C C O H H H H O H H H H H H H H Water H O H H H O O H H H C H N C N O C O –NH3 H C H H H H H Glycine 8 Figure 5. The Strecker synthesis of glycine, an amino acid. Reaction of the nucleophilic centers on the nitrogen of ammonia, the carbon of the cyanide anion and the oxygen of water with electrophilic centers on formaldehyde and key intermediates. The movements of pairs of electrons in the reactions are illustrated using curved arrows.

chemical formulas, the reaction is HCHO þ general. It can be used to prepare any amino HCN þ H2O¼C2H5NO2;thisequation“bal- acid for which the corresponding aldehyde is ances.” Ammonia is used in the first step, and is available, not just formaldehyde. For example, released in the last step. Therefore, ammonia is a if we start with (CH3CHO) rather catalyst for the reaction, being consumed and than formaldehyde, the amino acid formed in equal amounts in the reaction cycle. (CH3CH(NH2)COOH) is formed (Fig. 6). This sequence of reactions is known as the Analogous processes can be drawn for many Strecker synthesis of amino acids, named after of the amino acids found commonly in proteins the chemist who developed it in the 1860s. from terran organisms. The Strecker synthesis is driven by the innate The Strecker pathway accounts for some reactivity of nucleophiles and electrophiles, products found in Stanley Miller’s experiments and proceeds spontaneously and in reasonable attempting to reproduce early Earth’s atmos- yield. Further, the Strecker synthesis is quite phere (Miller 1955). Miller found that amino

Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 7 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

S.A. Benner et al.

O H Amide Acetaldehyde Imine C – H CN O H3C H H N C C H H CH +H O N C N NH –H2O 3 2 H H H H CH3 H Ammonia

O H +H O H 2 N C H C O –NH3 H CH3 Alanine Figure 6. Strecker synthesis of alanine starting from acetaldehyde.

acids were generated after electrical discharges ultraviolet light. The results obtained under a from electrodes were passed through an atmos- variety of conditions are complex red-brown phere of hydrogen, methane, and ammonia over mixturesoforganicmolecules,oftencalled“tho- water. The electrical discharge appears to have lins” (without making a distinction between generated the formaldehyde and cyanide needed what those mixtures contain) (Sagan et al. as precursors for Strecker syntheses. Once these 1978) (Sagan and Khare 1979). A partial inven- were formed, the synthesis of amino acids oc- tory of molecules comprising certain tholins is curred spontaneously. shown in Table 1. As the is The simulated atmosphere chosen by Miller also red-brown complex mixture of organics, it for his laboratory experiments was considered too may be said to contain “tholins,” although at the time to approximate the atmosphere of by ignoring the details of the mixture, this early Earth. Today, many models hold that the name is not particularly useful. However, there amount of methane on early Earth was much is little doubt that both formaldehyde and smaller than that used in the Miller experi- are key intermediates being ments. Instead, the carbon inventory of the generated from these atmospheres as well. early Earth is today modeled as being present largely as carbon dioxide (Kasting 1993). Curved Arrow Mechanisms for Forming Accordingly, nonbiological syntheses of bio- Nucleobases for RNA and DNA molecules under these conditions have been sought. Again, organic molecules are easy to Curved arrow mechanisms can be used to gener- get when carbon dioxide and water is subjected ate nonbiological routes for the synthesis of to electrical discharge, ionizing radiation, and many molecules in biology. For example, the

Table1. Some organic compounds identified in tholin mixtures. (Sagan et al., 1978; Sagan & Khare, 1979; Pietrogrande et al., 2001). Hydrogen sulfide Hexene Hydrogen cyanide Heptene Pyridine Ammonia Butadiene Styrene 2,3 Pentadiene Propane Toluene 2-Methylpyrimidine Butane Thiophene 4-Methylpyrmidine Ethene 2-Methyltiophene 3- Butenenitrile Methylmercaptan Butyne Butene Ethylmercaptan Pentene Propylmercaptan Carbon dioxide Carbon disulfide Methylisocyanate

8 Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

Planetary Organic Chemistry and the Origins of Biomolecules

Oro´-Orgel synthesis (Fig. 7) exploits the reactiv- unit. This, at neutral pH, loses an Hþ and 2 ity of HCN to make adenine (C5H5N5), one of becomes a COO carboxylate anion, a species the five nucleobases (symbolized as A, C, T, G, having a negative charge. This negative charge and U) used to store information in DNA and discourages any electron rich nucleophile from RNA. The cyanide anion again reacts as a nucle- attacking the C¼O unit of a carboxylic acid, ophile, thistimewithamoleculeof HCN serving making it rather stable under standard condi- as an electrophile. The combination of reactions tions. The other bonds in an amino acid are was studied in the laboratory Sanchez, Ferris, single bonds, including C-C, C-H, and C-N and Orgel, and provides one of the canonical bonds. These are, as noted earlier, also rather examples of nucleophile-electrophile chemistry stable. This means that amino acids are quite (supplemented with some photochemistry) in metastable, a feature of their reactivity that the prebiotic chemistry literature (Oro´ 1960; undoubtedly accounts for their presence in Sanchez et al. 1967). . Metastability is a more serious issue with respect to carbohydrates. Many reaction schemes Curved Arrow Mechanisms for Forming might give carbohydrates under prebiotic condi- Carbohydrates tions. However, carbohydrates are often unstable Amino acids are “easy” components of modern under the conditions where they are formed. Let terran from the perspective of us review just two examples where synthesis of metastability. The carboxyl group of an amino carbohydrates is possible, at least the laboratory, acid has a C¼O unit. Although this carbon is before turning to metastability as an issue. potentially an electrophilic center just as the For example, Eschenmoser exploited the C¼O unit of formaldehyde is, the C¼O unit reactivity C¼O aldehyde groups to generate of a carboxylic acid is also attached to an 2OH ribose, the carbohydrate found in RNA. His

H H H H N H N H C C HCN N C H NC C N H – CN – – CN CN

H H H H N N C N ν N C N ν C h h N C C H H C H H C H Cis-trans H C Rearrangement N C C N H N N C N N C isomerization H H

H H H H N H N N N C N C N C C N N C C C H N C C H C C H H N C C H N C C H N N H N N H H H H H Adenine Figure 7. HCN yields adenine via the Oro´-Orgel synthesis.

Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 9 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

S.A. Benner et al.

proposed reaction sequence begins with an Not quite ribose, but avery interesting prod- interesting starting material derived from HCN uct nevertheless. This process occurs in the labo- having two C-N single bonds in a three-mem- ratory when the starting materials are mixed bered ring (Fig. 8). Key in this process is the use with formaldehyde (Mueller et al. 1990). The of phosphate as a nucleophile to break one of starting material has not been detected naturally these C-N single bonds, replacing a C-N bond in the cosmos by microwave , but by a C-O bond. The breakability of the C-N this is hardly dispositive; the material would be bond isenhanced because itis strainedin athree- difficult to detect. membered ring. Another route for the prebiotic synthesis of The resulting intermediate is called a cyano- carbohydrates is based on the formose process, hydrin. It can fall apart to give an imine (a struc- described more than a century ago by the ture with a C¼N unit that we saw in the Strecker Russian chemist Aleksandr Butlerov (Butlerov synthesis). The imine can be hydrolyzed by 1861). If one incubates formaldehyde (HCHO) water to forms an aldehyde having an -O- in a hot solution of calcium hydroxide phosphate unit next to the C¼O unit; this is (Ca(OH)2, pH 12.5), nothing happens at first. called glycolaldehyde phosphate. This species However, after some time, the formaldehyde is can enolize and add formaldehyde to give gly- rapidly consumed. The mixture turns brown ceraldehyde (we will discuss enolization and and acquires a sweet taste; the product formose addition reactions more in a bit), which can smells like toasted marshmallows. Further, if serve as an electrophile to form a new bond to the reaction is stopped at the right time, some the enol of a second glycolaldehyde, to give five-carbon carbohydrates can be isolated from ribose-2,4-diphosphate as a major product. it, perhaps even ribose.

Phosphate O P HO O HO O O H2O HO P NH ν H HO O P 2 h CN HO O H2CC HO H C H CCH H 2 N CN 2 H2CC CN H NH2 NH Aziridine H “cyanohydrin” Imine H3N

Enediolate O H C = O H H C OPO3 C = H H C OPO3 O H C O H O H C H C OPO = C 3 Enediolate H C OPO = HO C H H C OPO = 3 3 = Aldol H C OH Aldol H C OPO3 H H addition C addition H H C OH O H

Figure 8. The Eschenmoser synthesis of ribose-2,4-diphosphate from a proposed starting material derived from HCN.

10 Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

Planetary Organic Chemistry and the Origins of Biomolecules

A half century ago, modern chemical analy- these compounds also have a C-H unit adjacent sis was brought to bear on the process by Ronald to the C¼O unit. From these, Hþ can be lost in Breslow (Breslow 1959). It involves the repeti- an enolization reaction, allowing them to react tion of two reactions shown in Figure 9: further as nucleophiles. Eventually, the products in the mixture undergo still more reactions that 1. removal of a proton (Hþ) from a carbon next are too many to capture in a single figure (and to a C¼O (carbonyl) group, known as eno- this is before considering ). Unfortu- lization, and nately, five carbon carbohydrates such as ribose, of particular interest for the RNA-first model for 2. attack of the resulting enediolate (a nucleo- the origin of life, are also metastable. They react phile) on HCHO (an electrophile) to form a further, especially at high pH. new carbon-carbon bond, known as an aldol addition. SOLVING THE METASTABILITY PROBLEM IN By repeating these reactions again and again, CARBOHYDRATE CHEMISTRY complex mixtures of organic species, all having a Appendages to Carbohydrates to Improve ratio of carbon:hydrogen:oxygen of 1:2:1 ( just their Metastability likethe starting formaldehyde)can begenerated. A subset of these reactions is shown in Figure 10. Two general classes of solution have been pro- Many popular books on the origin of life posedtoaddressthecomplexityandmetastability regard the formose process as a solution to the of carbohydrate formed by prebiotic processes. problem of prebiotic carbohydrate formation One proposes that free carbohydrates were never (Dyson 1985); some see it as a satisfactory formed on early Earth as prebiotically relevant approach for the prebiotic synthesis of ribose. species. In this class of solution, carbohydrates This is not the current consensus, however (Sha- are proposed to have entered early metabolism piro 2007). Both the complexity of the product joinedtoanotherchemical moiety thatimproves mixture(asshowninFig.10)andthemetastabil- their stability. One proposal is described by John ity of its components are at issue. Many of those Sutherland, who appends the emerging carbo- components themselves have C¼O units (and hydrate to an emerging nucleobase (Powner therefore can react as electrophiles). Many of et al. 2009).

H H H O O H O O H O O H C C CC HO O H R Enolization R′ R R′ H H H H O O H O O H O CC CC ′ Aldol addition R R R′ R R″ C R′′′ H R″ R′′′ C O H O H O HO

Figure 9. Two reaction types create formose. Top. In an enolization reaction, a base (here, hydroxide) removes a proton (Hþ) from a carbon adjacent to a C¼O unit to give an enediol. The enediol can then react as a nucleophile in an aldol addition reaction. Bottom. In an aldol addition reaction, an enediol reacts as a nucleophile to form a new bond to the C of a C¼O unit (the electrophilic center). If the carbonyl species is formaldehyde, R“¼R”’¼H.

Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 11 Downloaded from ..Bne tal. et Benner S.A. 12 http://cshperspectives.cshlp.org/ HOH2COH O HOH OH HO C O OH Formaldehyde O C C HO C O HO C OH C OH HCHO H O C2,3 C OH C C C C HO C OH 2 2 HH H H C H C C C Fe++ HH H HH H HH H HCHO HH HH HH OH HH HH Glycolaldehyde HH OH HH O C C OH C C OH C C C C OH HO C C HO C OH HO C C HO C C C4 OH HH HO H HO H H HO H HH Glycerol HCHO

HH OH HH HH O HH OH HH O HH OH

C5 C C C C C C HH HO CH2 C C OH C C OH HO C C OH HO C C OH HO C C HO C C onOctober3,2021-PublishedbyColdSpringHarborLaboratoryPress C C O HO H OH HO H HOH HO C C HO CH2 HH HO CH2 H HCHO HO H H OH OH

OH OH OH HH O HH HHH C OH HH HH OH HH HH HO CH ieti ril as article this Cite C6 C C C 2 2 HHH C OH H 2 HO C C OH C C C C C C C C O C C C HO C C HO C C OH HO CH HOH HO C C OH HO C C OH 2 HCHO HO CH HO CH H HO H O OH 2 OH 2 HO H OH OH OH OH OH OH C7 OH OH HH O HH HH HO CH HH HH HO CH H H H H C OH H HHH C OH HOH OH 2 2 2 2 HHH2C OH odSrn abPrpc Biol Perspect Harb Spring Cold C C C C C C C C C C C C OH C C C C C C OH HO C C OH HO C C OH HO C C OH HO C C O HO C C C HO C C HO CH OH HCHO HO CH HOCH HO CH O 2 HO H HO CH2 HO H O HH HO H OH HH 2 2 2 OH OH OH OH OH Enolization C8 OH OH OH OH OH Aldol at primary C HH H C OH HO CH HHH C OH O HH HO CH H HH HO CH HOH Aldol at secondary C 2 2 2 2 2 Retroaldol C C C OH C C C OH C C C C C C OH Bond broken in retroaldols HO C C C HO C C C HO C C O HO C C C Fenton reaction HO CH HO CH HO H O HH HO H HO CH2 HH 2 2 HO CH2 O HH OH OH OH OH

2010;2:a003467 Figure 10. This figure shows the complexity that is possible simply by repeating the two reactions shown in Figure 9. It shows the structures of organic molecules made of only carbon atoms (C), hydrogen atoms (H), and oxygen atoms (O) in a ratio of 1:2:1. The compounds are ordered by size, with compounds containing two and three carbon atoms (C2 and C3, respectively) at the top, and compounds containing four, five, six, seven and eight carbon atoms (C4, C5, C6, C7, and C8) ordered in rows below. The arrows show reactions that interconvert these compounds. The heavy black arrows show the addition of formaldehyde (HCHO) to a species in the row above; this converts that species to a new species with one more carbon atom. The open arrows show reactions that interconvert species having the same number of carbon atoms. Red arrows show reactions that fragment a larger molecule to give two smaller molecules, where the bond that is broken in the fragmentation is red. Blue compounds are dead-end compounds that accumulate in the reaction. A chemist must intervene to prevent this mixture from evolving further to give still more complexity. Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

Planetary Organic Chemistry and the Origins of Biomolecules

Eschenmoser’s proposal (Fig. 8) embodies a H2MoO4,H4GeO4, and H3AlO3, although similar idea. By introducing phosphate early in many of these acids are not observed, preferring the sequence, the intermediates are not glycolal- to precipitate or form complex ions. Silicic acid, dehyde, glyceraldehyde, and ribose, but rather for example, precipitates as SiO2, the mineral glycolaldehydephosphate,glyceraldehyde phos- known as quartz, in many forms, including phate, and ribose-2,4-diphosphate. The nega- gem opal. tive charges on the phosphate groups diminish Consider,forexample,boron.Standingright the rate at which Hþ is removed, therefore before carbon in the Periodic Table, boron has decreasing the rate of enolization. The phos- only three electrons in its outer shell. This means phates attached to these carbohydrates therefore that boron must pick up five more electrons decreases the reactivity of these species as nucle- to get the eight that it wants in its outer shell. ophiles, allowing them to survive longer under Oxygen atoms in water or the -OH hydroxyl conditions where they are formed. groups of organic molecules offer those elec- It remains to be seen whether attaching car- trons. Each of these has two unshared bohydrate fragments to these or other groups pairs of electrons. Therefore, boron (III) atoms increases metastability sufficiently to allow the bind to water oxygens. In geology, borate is an carbohydrate derivatives to be plausible prebi- anion whose central boron atom is bonded to otic participants in subsequent reactions pro- four oxygen atoms. posed in an RNA-first model for the origin of Boron bonds especially well to organic mol- life. It is clear, however, that the universe of ecules that have two adjacent hydroxyl groups, chemical structures is far from explored for pos- molecules known as 1,2-diols. Pentoses like sible attachments that might confer metastabil- riboseintheircyclicforms(Fig.11)are1,2-diols. ity with respect to decomposition and (still Their adjacent -OH groups bind borate tightly better) activation for the next steps in a useful (Fig. 12). Still better, their cyclic borate com- prebiotic sequence that does not require contin- plexes lack C¼O groups necessary for carbohy- ued human intervention. Much more work in drates to react to form tar. this direction is needed. As many have noted, the borate complex of ribose is stable, even at high pH. Indeed, for over 40 years, this stability has been used as part of Minerals to Improve the Metastability of synthetic procedures in the laboratory that Carbohydrates interconvert ribose and other five-carbon sug- A second line of thinking looks to the mineral ars, including ribulose and arabinose (Angyal world to preserve carbohydrates long enough 2001; Mendicino 1960). to allow them to accumulate to an extent useful To convert this common knowledge into a for them to participate in subsequent steps in a useful prebiotic hypothesis, two questions are prebiotic synthesis of Darwinian chemical relevant. The first asks whether minerals con- systems (Prieur 2001; Ricardo et al. 2004). The taining borate are likely to be available on prebi- role of mineral elements to create new reactivity otic Earth. As an element, boron is scarce in the has been reviewed by Robert Hazen. Here, we cosmos (Zhai and Shaw 1994), especially relative consider the opposite, the use of mineral ele- to silicon, carbon, and other familiar elements. ments to remove reactivity. Balancing this is the fact that boron is not a Several elements in mineral form are known good mineral-forming species. It therefore tends to bind to carbohydrates in a way that dimin- to be concentrated in residual igneous melts, ishes their reactivity. The best known of these where it appears on the surface. The predomi- are borate (Chapelle and Verchere 1988), silicate nant mineral species containing boron are tour- (Lambert et al. 2004), vanadate, molybdate, ger- malines. These are easily weathered in acid to manate, and (to a lesser extent) aluminate form borate salts, which are quite water-soluble. (Schilde et al. 1994). Purely formally, these are The fate of water-soluble borate salts derived from the acids H3BO3,H4SiO4,HVO3, depends on whether run-off streams deliver

Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 13 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

S.A. Benner et al.

Ribose Arabinose Xylose Lyxose open form open form open form open form H O H O H O H O C C C C H C OH HO C H H C OH HO C H H C OH H C OH HO C H HO C H H C OH H C OH H C OH H C OH H C OH H C OH H C OH H C OH H H H H

OH H OH OH H OH HO CH HO CH HO CH2 H H HO CH2 2 2 O O O O CH CH OH CH CH OH CH CH OH CH CH OH CC CC CC CC H HO OH HO H H OH H Ribose Arabinose Xylose Lyxose cyclic form cyclic form cyclic form cyclic form

H H H C H H H H OH H C H CH2OH O CH2OH C O O C O Ribulose H2C H2C Xylulose HO C H open form H C OH OH OH open form H C OH HO OH HO H H C OH H C OH Ribulose Xylulose H C OH H cyclic form cyclic form H Figure 11. Carbohydrates with 5 carbons (pentoses or pentuloses, penta¼5). Pentoses have a HC¼O group in their open form; pentuloses have a C¼O unit flanked by two carbons All have the formula C5H10O5. All can exist in both open and cyclic forms, as the C of the C¼O unit is an electrophile and can react with the O of an –OH group as a nucleophile to form a ring. In many cases, more than one cyclic form is possible. Different pentoses differ in how their atoms are arranged in space. We represent these 3D orientations on a 2D sheet of paper by drawing thicker bonds (which are forwards) and placing the H and OH groups up or down. Cyclic pyranose forms (with six atoms in the ring) are not shown.

them. If they run into the ocean, borate is diluted. However, if they end up in a dry basin such as Death Valley or the Dead Sea, borates

HO CH2 end up in minerals evaporated fromwater (evap- orites). Borax, a sodium borate salt, is one of H O these minerals, and is mined in Death Valley.

C H H C Other boron minerals found in Death Valley H O include ulexite, a sodium calcium borate min- CC eral, and colemanite, which is a calcium borate. Alkali is also easily generated from igneous B O O O rocks. One source for alkalinity in geology is H the mineral peridotite (its gem form is peridot). O This green mineral is a magnesium iron silicate, Figure 12. The NMR structure of the borate-ribose and is widespread in rocks called serpentines. complex, with boron complexing adjacent 1,2- When serpentines erode in water, serpentini- hydroxyl groups. zation occurs. Serpentinization is a process

14 Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

Planetary Organic Chemistry and the Origins of Biomolecules

that creates dihydrogen gas, the related reducing and (CH3OH) as the dead-end prod- power, reduced organic molecules, and magne- ucts. Even at 6 mM, borate significantly slows sium hydroxide, a base. For example, Lake formose cycling. Mono, just to the east of the Sierra Nevada Obviously, the ability of borate to stabilize mountains in California, has a pH as high as pentoses (such as ribose and arabinose) and 12, similar to the pH in the laboratory formose pentuloses (such as ribulose and xylulose) is reaction, all arising via serpentinization. useless if borate prevents their formation. The second question is more problematic. Thus, the second question asks whether proc- Simply adding borate does not solve problems esses exist that might form carbohydrates from with the prebiotic formose synthesis of carbohy- HCHO in the presence of borate minerals with- drates from formaldehyde. The synthesis begins out the slow reaction between two HCHO mol- with an exceptionally slow reaction between two ecules, and therefore without the need to have HCHO molecules, a reaction that creates just all of the cycles shown in Figure 10. single molecules of glycolaldehyde. The formose One particularly promising cycle that is process therefore relies on the complex reaction guided by borate was hypothesized (Fig. 13). cycles that use glycolaldehyde molecules (Hollis It starts with dihydroxyacetone, which lacks a 2000) to “fix” more HCHO to give highercarbo- 1,2-dihydroxyl unit. This should, according to hydrates (Fig. 10). Unfortunately, intermediates the hypothesis, enolize in the presence of borate in these cycles also have 1,2-dihydroxyl unitsthat and initiate a simple catalytic cycle to fix HCHO bind to borate. Accordingly borate (at 100 mM) (reactions within the red dotted box in Fig. prevents formose cycling, with HCHO dispro- 10). Proceeding clockwise in Figure 13, the portionating instead to give formate (HCOO2) enediol of dihydroxyacetone should react as a

HH O HH Xylulose + H H HH OH C C C HO C C OH ribulose 9:1 HO C OH C C OH (below) HH HO C C HOH HOH Threo HCHO C C O Cannot OH HH C5l addition HO C C enolize Borate inhibits H HH OH HCHO Branched HO H Cannot HOCH2 OH enolize HH O C C OH pentoses H H HO C C Addition H C OH C5b C OH C C C C OH HO HO C C Enolization HO H H HH O HH C C O HOH HH C4e HO C C HCHO HCHO Erythrulose Erythro addition H addition C4k HO H Dihydroxyacetone Retroaldol fragmentation slowed by borate complexation O OH H OH OH O HO C OH HCHO H C OH O C OH HO C H C C C C C C C HO Enolization C H C H HH HH Enolization Addition Enolization OH HH H H H H H Glyceraldehyde C2e Glycolaldehyde H O C3e 3 + 2 2 2 C3a addition Fe++ C2a H HO H H HO H H HO H H HO H H O H O H O H O Addition HO H C C C C C C C C C C C C HO C C HO C C HO C C HO C C HO C OH H H H H C C Borate controls HO H HO H HO H HO H HO H HO H HO H HO H diastereoselectivity HH HH Xylose Lyxose Ribose Arabinose HO H Glycerol Pentoses, borate stabilizes O HO C C Addition C C H 3 + 2 Equilibrate in borate addition HO H O HO H O H H HO H HO C C OH HO C C OH Threose, C C C C C C borate stabilizes HH H OH HH HH H OH HH Xylulose Ribulose Pentuloses, borate stabilizes Figure 13. Borate-guided prebiotic path to form carbohydrates from formaldehyde. Compounds in green are known to be prebiotic. Compounds in blue cannot enolize. Dihydroxyacetone is shown in brown.

Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 15 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

S.A. Benner et al.

nucleophile with HCHO (C1) to give erythru- give the 1,2-enediol (C4e), which should fix a lose (C4 k,3þ 1 ¼ 4). third HCHO to give either the linear or Erythrulose cannot form a cyclic hemiacetal branched C5 species at the top (C5l) and top- as do pentoses (Fig. 11), and therefore does not right (C5b) of Figure 13 by at bind borate tightly. It can, however bind borate the less or more hindered enediol carbon of less tightly through its 3,4-diol unit (Fig. 14). As C4e (these are both 4 þ 1 ¼ 5 reactions). the boron atom in the complex carries a nega- ThebranchedpentosesC5bdonothaveeno- tive charge, this binding should direct the eno- lizable , and therefore cannot react as lization of erythrulose away from the borate to nucleophiles. They can, however, cyclize to give

HCHO reacts as electrophile at nucleophilic sites H H OH H H OH

C C OH C C OH HO C C O C C

OH H H HO B O H H Loss of Borate inhibits Loss of Enolization HO red H enolization on red H to the left left side

H H O H H O

C C OH C C OH HO C C O C C

HO H H H HO B O H H H Erythrulose OH Loss of Loss of Enolization blue H blue H to the right

H H OH H H OH

C C OH C C OH HO C 1C O C C

HO H H HO B O H H OH HCHO reacts as electrophile at nucleophilic sites OH

H H O H H H H HO CH2 C C C C C O 1 HO C C OH HO C C

HO H H OH HO H H

Cannot cyclize, Cyclizes, reacts further binds borate, is stabilized H H H H OH OH OH C C O H O H H2C C H2C C CCOH CCOH

HO OH OH Borate Borate

Figure 14. Binding of borate to the 3,4-diol unit of erythrulose should direct enolization to give the 1,2-enediolate, as this removes the Hþ from the carbon the farthest from the negative charge on boron.

16 Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

Planetary Organic Chemistry and the Origins of Biomolecules

species that have adjacent 1,2-diol units that can hemiacetals that coordinate borate, threose, bind borate. Therefore, their complexes with and erythrose are also quite metastable in borate borate should accumulate. The uncomplexed buffers at pH 10.4. branched pentoses C5b species can, however, The same process can be observed with sili- undergo retroaldol fragmentation to generate cate. Silicate dissolves as its sodium salt only at C2a and C3e (Fig. 13, 5 ¼ 2 þ 3). The C3e is, high pH; at lower pH, silica precipitates. Never- of course the very same enediol that was gener- theless, a solution of concentrated sodium sili- ated from dihydroxyacetone, and can start the cate (about 20%, “water glass”) has a pH of cycle anew. The enediol of newly formed glyco- 11.5to12.0.Evenat100:1folddilution,however, laldehyde (C2e) should react with HCHO (C1) glycolaldehyde and glyceraldehyde react to form to form C3a (a 2 þ 1 ¼ 3 reaction), which is arabinose and other pentoses, which are also sta- hypothesized to enolize to form enediol C3e, bilized by complexation, this time to silicate again the same enediol that is formed byenoliza- (Lambert et al. 2004). tion of dihydroxyacetone. Thus, we can go To complete the metabolic cycle, we begin around the cycle again, fixing more HCHO. by recognizing that the branched pentoses A rich collection of labeling experiments with C5b are stabilized by borate, even though they H13CHO and 1-13C-glycolaldehyde show that suffer retroaldol fragmentations rapidly in the each of these steps is possible under conditions absence of borate (half life of ca. 30 minutes at that might have occurred on early Earth (Benner, 65 8C). The half-life for the retroaldol fragmen- Kim, and Kim unpublished). These studies often tation is ca. 24 hour at 40 mM borate, and use alkaline borate solutions buffered to pH 9 to becomes still longer at still higher borate con- 11 by carbonate, which is available from atmos- centrations. Thus, on a prebiotic Earth in the pheric carbon dioxide. Formation of 2’-13Cand presence of 200 mM borate, branched pento- 4-13C 2-hydroxymethylerythrose (C5b), 2’-13C- ses formed from HCHO and glycolaldehyde 2-hydroxymethylthreose (C5b), and linear and/or glycerol (Cooper, 2001) should have 1-13C -1,2,4,5-tetrahydroxypentan-3-one (C5l) formed stable borate minerals. These, in turn, from H13CHO proceeds easily without human could have been reservoirs for subsequent steps intervention in these buffers, afact that was estab- in the prebiotic synthesis of RNA. lished by comparing carbon-13 nmr signals in Of course, they could also support the cycle the products with the signals of labeled authentic by slow fragmentation, or by fragmentation material obtained by direct synthesis. after they are stripped of borate by another car- These steps in the presence of borate can be bohydrate (e.g. ribose) that binds borate more examined individually to determine how robust tightly. Todemonstrate the cycle, C5b was incu- they are with respect to changes in conditions. bated with H13CHO at 20 mM borate. The mix- For example, if the concentration of glycolalde- ture gives rise to labeled C5b as well as other hyde is present in two-fold excess over HCHO, products that can arise only if C5b had suffered addition of H13CHO to glycolaldehyde (1 þ 2 a retroaldol fragmentation to give glycolalde- ¼ 3, at 658C) followed by reaction with a second hyde and glyceraldehyde. This establishes the moleculeofglycolaldehyde(ina3 þ 2 ¼ 5reac- complete cycle. tion) gave 5-13C-ribose, 5-13C-arabinose, and These results show the possibility of a for- 1-13C-xylulose as major metastable products. maldehyde fixation cycle, where C2 and C3 add Reaction of glycolaldehyde (C2e) in the absence HCHO to give branched speciesthat eitheraccu- of CH2O gave four-carbon sugars, threose, and mulate or suffer retroaldol fragmentation to give erythrose in 2 þ 2 ¼ 4 aldol addition reactions. glycolaldehyde and glyceraldehyde. These can The formation of threose is significant because either capture HCHO to go around the cycle this tetrose can replace ribose in the backbone again, or can combine to give xylulose and ribu- of RNA-like genetic molecules (Scho¨ning et al. lose as stable borate complexes. 2000; Ebert et al. 2008; Horhota et al. 2005). Under kinetic control, the ratio of C5 spe- Consistent with their ability to form a cyclic cies formed depends on the relative rates of

Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 17 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

S.A. Benner et al.

aldol additions and fragmentations, the con- may have emerged from the hypothetical cycle centrations of various species, pH and borate in Figure 13. One leakage product is 1,2,4,5- concentration. However, under prolonged tetrahydroxypentan-3-one, shown the top of time on a primitive Earth, thermodynamics Figure 13 (C5l). It is a linear species formed by might well have determined the final products the addition of HCHO to the erythrulose ene- rather than kinetics. Equilibration within two diol (C4e) at its less hindered center (4 þ 1 ¼ sets of stereosimilar five-carbon species (ribu- 5). This species cannot form a cyclic hemiacetal lose, ribose and arabinose, and xylulose, xylose, to bind borate; it therefore reacts further in and lyxose) is well known under a variety of carbonate-borate buffer. mechanisms (Angyal 2001). Thus, a route that Remarkably, the leakage is “productive.” gives ribulose in the presence of borate also gives With a half-life of 4-5 days, synthetic 1-13C-C5l ribose, whereas a route that gives xylulose in the in carbonate-borate buffer gives 1- and presence of borate also gives xylose. More 5-13C-xylulose (90%) and 1- and 5-13C-ribulose slowly, the two pentose-pentulose sets them- (10%).Ribuloseisknowntobeisomerizedinthe selves interconvert (Fedoronki and Linek presenceofboratetoribose.Thus,leakagebythis 1967). Absent borate, the equilibrium ratios route forms the products that one wants, but in a can be represented as a series (xylose 1:1 xylu- different way from the aldol addition of a glyco- lose 85:15 ribulose 1:9 arabinose 4:1 ribose) laldehyde and glyceraldehyde. (Sultana et al. 2003), with the ratio of any two C5 carbohydrates calculable by multiplying the appropriate ratios (thus the xylulose:ribose More Mineralogy to Manage Carbohydrate equilibrium is 5:2). Borate favors the -uloses Metastability in each set over the -oses; among the pentoses, borate complexation favors ribose (Mendicino As before, borate’s ability to stabilize carbohy- 1960; Li et al. 2005). drates offers both good news and bad news. In particular, retroaldol fragmentation of branched pentoses C5b is slowed dramatically by borate. Challenges of Prebiotic Metabolic Cycling Prebiotic chemical scenarios often assume mil- Interestingly, borate-moderated cycles such as lions of years of time to manage slow reactions those in Figure 13 share some features proposed (a thermodynamic control scenario), and we for cycles hypothesized for “metabolism first” can easily imagine scenarios where the accumula- models (Shapiro 2007). Although the “genetics tion of branched carbohydrate outruns the con- first” versus “metabolism first” models for the centration of borate, leaving some unbound origin of life on Earth are currently being pre- complexes to suffer retroaldol reaction to give sented as adversaries (Orgel 2008), no logic ribose, which extracts a borate from a branched compels them to be. It is nearly certain that pentose to permit the transformative process to chemical processes that might be likened to go to completion. metabolism occurred on Earth before genetics, Further, the relative scarcity of boron has providing components of whatever genetic sys- become an issue. For example, Hazen has sug- tem did first emerge. gested that borate minerals are “exotic,” in part However, a cycle will not operate if more because they are concentrated in pegmatites than one equivalent of material leaks from the whose contents may have required too much cycle before completion of the cycle produces time took to accumulate (Hazen et al 2009). It a second equivalent of cyclable material. Prebi- is difficult to know in this context what “too otic metabolic cycles can easily be defeated by much time” might be, as weatherable igneous leakage that allows enough material to escape borates should have appeared on the surface as from the cycle to prevent product accumulation. soon as the planetary material separated and Tounderstandleakage inmineral-controlled concentrated as long as there was aqueous cycles, we might focus on leakage products that rain, dry land, and dry valleys.

18 Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

Planetary Organic Chemistry and the Origins of Biomolecules

Accordingly, some geologists are more com- α Threo β fortable with silicate control than they are with OH OH borate control, even considering the fact that sil- H2C H2C HO O H HO O OH icates, although more abundant, are less soluble + in water than borate. Aluminates are also candi- OH H date coordinating species, also abundant but OH OH also rather poorly soluble (aluminate and sili- cate together form clay). Nevertheless, other mineral forming species H H H H are known to coordinate carbohydrates, includ- H H HO C ingvanadate,germanate,andmolybdate(Schilde HO C O HH O HH et al. 1994). These are also “exotic,” and are more O C C O C C C C OH C C OH oxidized with respect to the mantle than borate, silicate, and aluminate. Thus, they are not likely H HO H H HOH to have been present on a reducing early Earth C3 migration C2′ migration in abundance. They may, however, have been H H present in significant amounts at specific locales. H H H H Balancing their exotic nature is some very O C O HH O C O interesting chemistry produced by these min- HO C C C C OH C HH eral components, especially at moderate pH. HO 6þ C C For example, molybdate (Mo ) minerals at H H OH C OH neutral (or slightly acidic pH) catalyze a Bilik H H OH reaction, a rearrangement of a hydroxyl-car- Linear pentulose Ribulose bonyl compound to give an isomeric carbonyl- hydroxyl compound (Petrus et al. 2001). This Hydride migration can convert branched tetroses such as C4b into erythrulose and branched pentoses such Ribose + arabinose as C5b to linear pentoses (Fig. 15). In both Figure 15. The Bilik reaction uses molybdate minerals cases, “dead end” products might be returned to catalyze the stereospecific isomerization of carbohydrates such as the branched pentoses to give to the catalytic cycle or to biologically interest- linear pentuloses such as ribulose at neutral pH ing species. under mild conditions. Indeed, incubating branched pentose (C5b)in the presence of molybdate (65 8C, 24 hour) leads alkaline solution. Thus, coupled with the borate- to an equilibrium mixture of C5b starting material guided (or silicate-guided or aluminate-guided) and linear xylulose, with some linear pentulose prebiotic fixation of HCHO and fluctuating pH, 1,2,4,5-tetrahydroxypentan-3-one (Kim and Ben- Mo6þ can generate free ribose and xylose from ner, unpublished). The rearrangement is stereo- C5b branched pentoses. This model envisions a specific; threo branched pentose gives ribulose. borate-rich (or silicate-rich) alkaline stream As noted above, xylulose and ribulose equilibrate emerging from serpentinizing igneous rocks slowly with Mo6þ to give xylose and ribose. A sim- above the Earth’s surface, with formaldehyde ilarreactionwasobservedwithCaþþ as catalysts, arriving from the atmosphere above as the prod- but at high pH. uct of electrical discharge through a moist atmos- phere dominated by carbon dioxide. Under these conditions, complexes of borate or silicate are The Need for Dry Land, Incompletely Reducing formed with branched pentoses, ribulose and Conditions, and Incompatible Species xylulose, and their corresponding pentoses, Toexploitmolybdate,thepHmustfall.Thiscould arabinose, lyxose, ribose and xylose. happen under a carbon dioxide atmosphere The alkalinity of such a stream would be through dissolution of atmospheric CO2 in an buffered over time by the dissolution of

Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 19 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

S.A. Benner et al.

atmospheric carbon dioxide. As the pH drops, carbohydratesandothercompoundscontaining the rate of formation of carbohydrates via eno- C¼O bonds. Hydrogen sulfide attacks C¼O lization reactions slows, but so does their rate of groups, just like water. Further, the bisulfite 2 decomposition. At the same time, borate com- (HSO3 ) anion reacts with C¼O units to form plexes weaken and silicate precipitates, but reasonably stable bisulfite addition products. molybdenum species capable of supporting Bisulfite is formed from , which the Bilik reaction become more abundant. was undoubtedly produced on early volcanoes According to this model, these conditions on Earth. transform the branched carbohydrates to linear The reaction of carbohydrates under plausi- carbohydrates, which themselves form pentoses bly prebiotic conditions in the presence of these like ribose (Mendicino 1960). If the pH rose, the other species has only begun (Weber, 2001). ribose would again become bound to and stabi- Much more research is needed before a model lized by borate. Such pH fluctuations would can capture the entire space of variable contents help equilibrium to be reached. in away that suggests what mixtures are produc- Certain models for early Earth are not com- tive and what mixtures are nonproductive, for patible with this scenario, of course. For exam- prebiotic synthesis on a complex planet. Ulti- ple, this scenario could not easily operate on a mately, compatibility issues need to be captured planet that was entirely flooded by water, as in a table that includes both organic and inor- borate and alkali would be diluted into the plan- ganic species, including cations and anions that etary ocean whose oxidation state, pH, and areincompatiblewithsolubility,redoxpotential, mineral composition would be that of the ocean and reactivity. as a whole. Dry land is needed to permit the concentration of borate and borate-organic evaporite minerals for this model; alternatively, compartmentalization is needed within the REFERENCES ocean, such as near subsurface vents, to have Angyal SJ. 2001. The Lobry de Bruyn-Alberda van Ekenstein such cycles. Thus, if the “super-” being transformation and related reactions. Glycoscience observed as extrasolar planets are indeed “water 215:1–14. worlds,” this cycle would be more difficult to Benner SA. 2009. Life, the Universe and the Scientific Method. Gainesville FL, Foundation Press. 320 pp. operate, requiring sequestration in specific sub- Bernstein MP, Dworkin J, Sandford SA, Cooper GW, oceanic environments. Allamandola LJ. 2002. Racemic amino acids from the Given land surfaces, however, the model ultraviolet photolysis of interstellar ice analogues. Nature makes sense. Fluctuating pH is not implausible, 416:401–403. Blackmond DG. 2010. The origin of biological homochira- as the high pH emerging from serpentinizing lity. Cold Spring Harb Perspect Biol 2: a002147. rocks is buffered beneath a CO2 atmosphere. Breslow R. 1959. On the mechanism of the formose reaction. Borate is readily delivered to alkaline washes Tetrahedron Lett 21:22–26. through the erosion of tourmalines which were Butlerov A. 1861. Bildung einer zuckerartigen Substanz almost certainly present in igneous rockon early durch Synthese. Annalen Chemie 120:295–298. 6þ Cairns-Smith Alexander. 1982. Genetic Takeover and the Earth, as they are today. Mo is an oxidized Mineral Origins of Life. Cambridge UK: Cambridge form of the element, but not a stronglyoxidizing University Press. 488 pages. 3þ þ 11 13 form (MoO2 þ H2Oþ 2Fe ¼MoO3 þ 2H þ 2 Chapelle S, Verchere J-F.1988. A B and C NMR determi- 2þ nation of the structures of borate complexes of pentoses Fe at þ236 mV; 2 MoO2 þ SiO2¼Si þ 2 and related species. Tetrahedron 44:4469–4482. MoO at 20.145 mV) Thus, that the formation 3 Cleaves HJ. 2008. The prebiotic geochemistry of formalde- of MoO3, which is hydrated to give molybdate, hyde. Precambrian Res 164: 111–118. would have been favored at neutral when the fer- Cooper G, Novelle K, Belisle W, Sarinana J, Brabham K, ric:ferrous ratio is one part in ca 1010. Garrel L. 2001. Carbonaceous meteorites as a source of sugar-related organic compounds for the early earth. Of course, other elements may be involved Nature 414: 879–884. in such processes. Sulfur is an interesting exam- Dyson F.1985. Origin of life. Cambridge University Press. 94 ple. Species containing sulfur interact with pp.

20 Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

Planetary Organic Chemistry and the Origins of Biomolecules

Ebert MO, Mang C, Krishnamurthy R, Eschenmoser A, Jaun Pinto JP, Gladstone GR, Yung YL. 1980, Photochemical B. 2008. The structure of a TNA-TNA complex in production of formaldehyde in Earth’s primitive atmos- solution: NMR study of the octamer duplex derived phere. Science 210:183–184. from a-(L -threofuranosyl-(3’-2’-CGAATTCG). JAm Pizzarello S, Shock E. 2010. The organic composition of car- Chem Soc 130: 15105–15115. bonaceous meteorites: The evolutionary story ahead of Fedoronki M, Linek K. 1967. Transformation of pentoses in biochemistry. Cold Spring Harb Perspect Biol 2: a002105. pyridine. Coll Czech Chem Comm 32: 2177. Powner MW, Gerland B, Sutherland JD. 2009. Synthesis of Hazen RM, Ewing RJ, Sverjensky DA. 2009. The evolution of activated ribonucleotides in prebiotically uranium and thorium minerals. Am Mineral 94: 1293– plausible conditions. Nature 459: 239–242. 1311. Prieur BE. E´ tude de l’activite´ pre´biotique potentielle de l’acide borique. C R Acad Sci Paris Chimie 4:667–670. Hollis JM, Lovas FJ, Jewell PR. 2000. Interstellar glycolalde- hyde: The first sugar. Astrophys J 540: L107–L110 Part 2. Sagan C, Khare NB. 1979. Tholins: Organic chemistry of interstellar grains and gas. Nature 277:102–107. Horhota A, Zou K, Ichida JK, Yu B, McLaughlin LW,Szostak JW, Chaput JC. 2005. Kinetic analysis of an efficient Sagan C, Khare NB, Bandurski LE, Batholomew N. 1978. DNA-dependent TNA polymerase. J Am Chem Soc 127: Ultraviolet-photoproduced organic solids synthesized under simulated jovian conditions: Molecular analysis. 7427. Science 199: 1199–1201. Joyce GF, Orgel LE. 1999. Prospects for understanding the Sanchez RA, Ferris JP, Orgel LE. 1967. Studies in prebiotic origin of the RNA world. In The RNA world 2nd edn. synthesis II. Synthesis of purine precursors and amino (eds Gestland R. et al.). Cold Spring Harbor Press Cold acids from aqueous hydrogen cyanide J Mol Biol Spring Harbor NY. 30:223–253. Kasting JF.1993. Earth’s early atmosphere. Science 259:920– Schilde U, Kraudelt H, Uhlemann E. 1994. Separation of the 925. oxoanions of germanium, tin, arsenic, antimony, tellu- Lambert JB, Lu G, Singer SR, Kolb VM. 2004. Silicate rium, molybdenum and tungsten with a special chelating complexes of sugars in aqueous solution. J Am Chem resin containing methylaminoglucitol groups. Reactive Soc 126:9611–9625. Polymers 22:101–106. Larralde R, Robertson MP, Miller SL. 1995. Rates of de- Scho¨ning KU, Scholz P,Guntha S, Wu X, Krishnamurthy R, composition of ribose and other sugars, implications Eschenmoser A. 2000. Chemical Etiology of nucleic acid for chemical evolution. Proc Natl Acad Sci 92:8158–8160. structure. The k-threofuranosyl-(3’!2’) oligonucleotide system. Science 290:1347–1351. Li Q, Ricardo A, Benner SA, Winefordner JD, Powell DH. Desorption/ on porous silicon mass spec- Shapiro R. 1987. Origins: A skeptic’s guide to the creation of trometry studies on pentose-borate complexes. Anal life on earth. Bantam Books, New York. Chem 77: 4503. 2005. Shapiro R. 1988. Prebiotic ribose synthesis. A critical analy- sis. Origins Life Evol Biosphere 18:71–85. Mendicino JF. 1960. Effect of borate on the alkali-catalyzed isomerization of sugars. J Am Chem Soc 82:4975. Shapiro R. June 2007. A simpler origin for life. Scientific American 296:46. Miller SL. 1955. Production of some organic compounds under possible primitive earth conditions. J Am Chem Sultana I, Mizanur RMD, Takeshita K, TakadaG, Izumori K. Soc 77:2351–2361. 2003. Direct production of D-arabinose from D-xylose by a coupling reaction using D-xylose isomerase D-tagatose Mueller D, Pitsch S, Kittaka A, Wagner E, Wintner CE, 3-epimerase and D-arabinose isomerase. J Biosci Bioeng Eschenmoser A. 1990. Chemistry of alpha-aminonitriles 95:342–347. Aldomerization of glycolaldehyde phosphate to Weber AL. 2001a. The sugar model. Catalytic flow reactor rac-hexose 2 4 6-triphosphates and. in presence of form- dynamics of pyruvaldehyde synthesis from triose cata- aldehyde. rac-pentose 2 4-diphosphates: rac-allose 2 4 6- lyzed by poly-D-lysine contained in a dialyzer. Orig Life triphosphate and rac-ribose 2 4-diphosphate are the Evol Biosph 31:231–240. main reaction products. Helv Chim Acta 73:1410–1468. WeberAL. 2001b. The sugar model. Catalysis by and Orgel LE. 2008. The implausibility of metabolic cycles on amino acid products. Orig Life Evol Biosph 31: 71–86. the prebiotic Earth. PLOS Biol 6: e18. Weber AL. 2007. The sugar model: Autocatalytic activity of Oro´ J. 1960. Synthesis of adenine from cyanide. the triose-ammonia reaction. Orig Life Evol Biosphere 37: Biochem Biophys Res Commun 2:407–412. 105–111. Petrus L, Petrusova´ M, Hricovı´niova´ Z. 2001. The Bilik Zhai M, Shaw DM. 1994. Boron Part I: reaction. Glycoscience 215:15–41. Boron in meteorites. Meteoritics 29:607–615.

Cite this article as Cold Spring Harb Perspect Biol 2010;2:a003467 21 Downloaded from http://cshperspectives.cshlp.org/ on October 3, 2021 - Published by Cold Spring Harbor Laboratory Press

Planetary Organic Chemistry and the Origins of Biomolecules

Steven A. Benner, Hyo-Joong Kim, Myung-Jung Kim and Alonso Ricardo

Cold Spring Harb Perspect Biol 2010; doi: 10.1101/cshperspect.a003467 originally published online May 26, 2010

Subject Collection The Origins of Life

Constructing Partial Models of Cells The Hadean-Archaean Environment Norikazu Ichihashi, Tomoaki Matsuura, Hiroshi Kita, Norman H. Sleep et al. Ribonucleotides An Origin of John D. Sutherland Christopher P. McKay Deep Phylogeny−−How a Tree Can Help Primitive Genetic Polymers Characterize Early Life on Earth Aaron E. Engelhart and Nicholas V. Hud Eric A. Gaucher, James T. Kratzer and Ryan N. Randall Cosmic Carbon Chemistry: From the Interstellar Membrane Transport in Primitive Cells Medium to the Early Earth Sheref S. Mansy Pascale Ehrenfreund and Jan Cami Origin and Evolution of the Ribosome The Origins of Cellular Life George E. Fox Jason P. Schrum, Ting F. Zhu and Jack W. Szostak Planetary Organic Chemistry and the Origins of From Self-Assembled Vesicles to Protocells Biomolecules Irene A. Chen and Peter Walde Steven A. Benner, Hyo-Joong Kim, Myung-Jung Kim, et al. Mineral Surfaces, Geochemical Complexities, and The Origin of Biological the Origins of Life Donna G. Blackmond Robert M. Hazen and Dimitri A. Sverjensky Historical Development of Origins Research Earth's Earliest Atmospheres Antonio Lazcano Kevin Zahnle, Laura Schaefer and Bruce Fegley

For additional articles in this collection, see http://cshperspectives.cshlp.org/cgi/collection/

Copyright © 2010 Cold Spring Harbor Laboratory Press; all rights reserved