Volume 14, Number 3, 2014 ª Mary Ann Liebert, Inc. DOI: 10.1089/ast.2014.1140

The Fuel Model of : A New Approach to Origin-of- Simulations

Laura M. Barge,1,2 Terence P. Kee,3 Ivria J. Doloboff,1 Joshua M.P. Hampton,3 Mohammed Ismail,4 Mohamed Pourkashanian,4 John Zeytounian,5 Marc M. Baum,6 John A. Moss,6 Chung-Kuang Lin,1 Richard D. Kidd,1 and Isik Kanik1

Abstract

In this paper, we discuss how prebiotic geo-electrochemical systems can be modeled as a and how laboratory simulations of the origin of life in general can benefit from this systems-led approach. As a specific example, the components of what we have termed the ‘‘prebiotic fuel cell’’ (PFC) that operates at a putative Hadean are detailed, and we used electrochemical analysis techniques and exchange membrane (PEM) fuel cell components to test the properties of this PFC and other geo-electrochemical systems, the results of which are reported here. The modular of fuel cells makes them ideal for creating geo- electrochemical reactors with which to simulate hydrothermal systems on wet rocky and characterize the energetic properties of the seafloor/hydrothermal interface. That electrochemical techniques should be applied to simulating the origin of life follows from the recognition of the fuel cell–like properties of prebiotic chemical systems and the earliest . Conducting this type of laboratory simulation of the emergence of will not only be informative in the context of the origin of life on but may help in understanding whether life might emerge in similar environments on other worlds. Key Words: Astrobiology— Bioenergetics— sulfides—Origin of life—Prebiotic . Astrobiology 14, 254–270.

1. Introduction membrane that induce a pH gradient between the membrane interior and exterior (i.e., the proton motive force) drive the dentifying the specific processes responsible for the fundamental generation mechanism for life (Mitch- Iemergence of life on Earth and then convincingly simu- ell, 1961; Jagendorf and Uribe, 1966; Harold, 1986). As in lating these same processes in the laboratory are among the life today, the contained within these most intriguing challenges for contemporary biogeochemists. and proton gradients is transduced into useful A large part of the difficulty is that plausible scenarios for the and stored in metastable energy currency origin of life must not only account for the fundamental such as (ATP) or inor- properties of life today but also for the geochemical condi- ganic pyrophosphate (PPi). It has been argued that the very tions on early Earth through which life arose. These condi- first were autotrophic; that is, they gained tions are such that we only have patchy knowledge of them, and assembled their building constituents from the simple and they present challenging engineering problems during molecules available to them on early Earth (e.g., H2,CO2, þ 2 - attempts at simulation. CH4,NH3=NH2 , HPO4 , and HS /H2S) (Fuchs, 1989, A primary process common to all extant life is the con- 2011; Berg et al., 2010; Say and Fuchs, 2010). The electron version of a pH gradient into energy-storing an- transfer involved in the generation of the proton hydride bonds via across a semipermeable motive force—many of which host inorganic sulfide– Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. membrane. In prokaryotes, as well as in mitochondria, the containing active centers—have been suggested to be ‘‘power plant’’ of , electric potentials across the among the most ancient catalysts (Eck and Dayhoff, 1966;

1Jet Propulsion Laboratory, Institute of Technology, Pasadena, California, USA. 2Blue Marble Space Institute of Science, Seattle, Washington, USA. 3School of Chemistry, University of Leeds, Leeds, UK. 4Centre for Computational Fluid Dynamics, University of Leeds, Leeds, UK. 5Molecular and Computational Program, University of Southern California, Los Angeles, California, USA. 6Department of Chemistry, Crest Institute of Science, Pasadena, California, USA.

254 THE FUEL CELL MODEL OF ABIOGENESIS 255

Baymann et al., 2003; Volbeda and Fontecilla-Camps, the last universal common ancestor (Gogarten et al., 1989; 2006; Vignais and Billoud, 2007; Lill and Siegbahn, 2009; Martin and Russell, 2007; Mulkidjanian et al., 2007; McGlynn et al., 2009; Nitschke et al., 2013; Schoepp- Schoepp-Cothenet et al., 2013). Because we consider that Cothenet et al., 2013). LUCA also had these basic fuel cell properties (Russell and Microbes are, therefore, essentially performing the same Hall, 1997; Nitschke and Russell, 2009; cf. Huang et al., chemical processes as those in -exchange mem- 2012; Lane and Martin, 2012), we suggest that some key brane fuel cells (Mitchell, 1977; Barbir, 2005), and like fuel steps at the origin of may then be simulated cells, biology uses proton and ion gradients to generate experimentally within the context of the electrochemical energy. In fact, recognition of this similarity within the fuel gradient architecture of a fuel cell. As life is an example of a cell has helped to implement the use of mi- far-from-equilibrium , which presumably crobes, their redox-active , and even mitochondria emerged from a far-from-equilibrium geological and geo- themselves as components of for biofuel cells due chemical system, it seems logical that a far-from-equilibrium to their excellent catalytic ability to transfer chemical system such as a fuel cell may provide an effective and promote environmentally significant redox reactions link between the two. (Arnold and Rechnitz, 1980; Heller, 1992; Chang et al., We set out to examine whether the fuel cell–like prop- 2006; Arechederra and Minteer, 2008; Huang et al., 2012; erties of certain geochemical environments, such as seafloor Tran and Barber, 2012). Certain geochemical environments interfaces and hydrothermal vent systems, could be simu- also constitute fuel cell–like systems, for example, at hy- lated in out-of-equilibrium electrochemical experiments and drothermal vents where electrical and pH potentials are ultimately modeled in the laboratory as a fuel cell. In a generated at the interface between reduced hydrothermal geological/geochemical fuel cell scenario, some or fluid and oxidizing seawater (Baross and Hoffman, 1985; other inorganic component would be required to act as an Russell and Hall, 1997, 2006; Martin and Russell, 2007; electrocatalyst as well as an ion transfer ‘‘membrane’’ that Yamamoto et al., 2013). In the hydrothermal vent example, separates contrasting chemical reservoirs. Depending on the the ambient geochemical potentials can be maintained scenario, this component could be a conductive mineral, gel, and mediated across physical boundaries, for example, or other porous material. For our preliminary experiments, electrically conductive biofilms, sediments, or an inorganic we chose to focus on (i) Fe sulfides and (ii) Fe-Ni phases chimney precipitate (Ludwig et al., 2006; El-Naggar et al., within iron . Both of these are viewed as be- 2010; Nakamura et al., 2010a, 2010b; Yamamoto et al., ing potential catalysts relevant for proto- 2013). A second geological example of fuel cell–like be- metabolic reactions and are considered to have been read- havior that is believed to have been of significance within ily available and accessible on early Earth (Williams, 1961,

early Earth environments is galvanic of iron- 1965; Eck and Dayhoff, 1966; Russell et al., 1994; Bryant alloys (associated with meteoritic infall) from contact et al., 2009, 2013; Nitschke and Russell, 2011). Metal sul- with seawater, which can in principle establish pH/Eh gra- fides in particular constitute an important part of the envi- dients (Bryant et al., 2013). ronmental fuel cell in modern hydrothermal vent systems as Because life universally depends on pH/Eh gradients to they facilitate transfer of electrons (Nakamura et al., 2010b). drive metabolism, geological environments that already The nature of mineral precipitation in a gradient is also generate similar electrochemical energy gradients are of significant, as the chemical disequilibrium can result in interest for origin-of-life studies. One such example is hy- complex precipitate structures such as hydrothermal ‘‘che- drothermal vents, which can be produced by magmatic ac- mical garden’’ chimneys (Haymon et al., 1983; Russell tivity (e.g., black smokers) or - chemistry (e.g., et al., 1989, 1994; Ludwig et al., 2006) that may exhibit -hosted alkaline vents). Alkaline hydrothermal electrochemical and ion-transfer capabilities (Ayalon, 1984; vents produced by serpentinization (the process of / van Oss, 1984). To model these conditions in the laboratory, oxidation via seawater interaction in a series of an iron sulfide electrocatalyst material was synthesized exothermic reactions) are of special interest to the origin of by separating an acidic, Fe2 + -containing solution and an al- life, as they can generate a very alkaline hydrothermal ef- kaline, sulfide-containing solution across a permeable bar- fluent and, thus, an ambient pH gradient in addition to an rier to form a self-assembling precipitate membrane in a electrical potential difference across the fluid interface gradient. This simulated geological precipitate separated (Baross and Hoffman, 1985; Russell and Hall, 2006; Martin two contrasting solutions and could mediate gradients, as et al., 2008; Russell et al., 2010; Lane and Martin, 2012). might occur in a natural far-from-equilibrium system such Serpentinization is likely to have progressed on early Earth as chimney growth at a hydrothermal vent. Electrochemical (Russell et al., 1989; Sleep et al., 2011; Arndt and Nisbet, experiments also were conducted by using geological ana- 2012) and may have been active elsewhere in the Solar log materials (both synthetic and field samples) as catalysts, Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. System, for example, on early , , or other icy and a proton exchange membrane (PEM) fuel cell apparatus moons with a water-rock interface (Vance et al., 2007; was constructed to simulate an electrochemically active Ehlmann et al., 2010; Wray and Ehlmann, 2011; Michalski geochemical interface. The fuel cell represents a holistic et al., 2013), possibly providing an environment for emer- chemical system that is well understood, amenable to gent prebiotic chemistry. computational modeling, and open to sophisticated analyti- The operations of extant life are analogous to those of a cal diagnostics. This systems approach has not yet been fuel cell, and some version of the fundamental components exploited experimentally within the sphere of abiogene- that make the biological fuel cell (ATP-synthase, sis and offers a potentially powerful theoretical and exper- ion-selective membranes maintaining pH/electrical gradients, imental model through which to explore such emergent the ) was also likely present in LUCA, physicochemical systems. These experimental techniques 256 BARGE ET AL.

can be applied to a variety of geochemical systems or analog activity or by water-rock chemistry. High- hy- materials and may even be used to simulate energetic pro- drothermal alteration on oceanic spreading ridges can pro- cesses at seafloor interfaces on other planets as well as to duce a hot ( > 350C) acidic hydrothermal fluid driven by determine whether prebiotic chemistry would be possible on magmatic intrusion that fuels black smokers, named after other wet, rocky worlds. the black particles that form as the rapid temperature/pH change induces precipitation of metal sulfides at the fluid interface. In the absence of magmatic heating, reducing 2. Examples of Fuel Cell-like Systems hydrothermal fluids can also be produced by serpentiniza- in Biology and tion, a process in which seawater permeates through frac- To envision experimental ways by which to test prebiotic tures in the Fe/Mg-silicate ocean crust and produces an chemistry within an electrochemical framework, it is useful alkaline, H2- and CH4-enriched, moderately hot (*100– to define and constrain the functional parts of the natural 200C) hydrothermal fluid (Russell et al., 1989; Kelley fuel cells of geology and biology, as well as a proposed et al., 2001, 2005; Bradley and Summons, 2010; Charlou prebiotic fuel cell (PFC) that may have powered inorganic et al., 2010; Klein et al., 2013), even in low-temperature and organic redox reactions on early Earth. exhalations (Neal and Stanger, 1983; Coveney et al., 1987; Fuel cells are primarily composed of an (the sur- Etiope and Sherwood-Lollar, 2013; Etiope et al., 2013). face where oxidation occurs), a (the surface where Both black smoker and alkaline vents can produce chimney reduction occurs), and an ion-conducting electrolyte that precipitates where hydrothermal fluids feed back into the physically separates reservoirs containing the fuel and oxi- ocean, and these naturally occurring mineral precipitates are dant (Haile, 2003a, 2003b). The physical setup of the fuel significant in that they provide a semipermeable and semi- cell that separates two chemically distinct reservoirs is conducting barrier between the two contrasting fluids. The essential, for it is this chemical/redox disequilibrium— redox-active fluid interface of a hydrothermal vent consists chemical gradients maintained over some distance—from of fuel/oxidant reservoirs separated by an inorganic mineral which electricity is ultimately produced. The essential precipitate and constitutes a naturally occurring fuel cell that components include the electrolyte, which physically sep- does not depend on biological processes to function. The arates the fuels and oxidants, allowing a voltage gradient to inside and outside surfaces of the hydrothermal chimney be maintained; the electrodes, which facilitate oxidation/ could be considered as anode and cathode, respectively. reduction and contain catalytic materials; and the fuel Depending on composition and environmental conditions, () and oxidant (), which the chimney itself may be able to catalyze the oxidation of transfer electrons to and from the electrodes. hydrothermal fuels and conduct resulting electrons to sea-

Mitochondria, the free energy factories of cells, water oxidants. The fuel cell–like properties of hydrother- are very efficient electrochemically, and they and their as- mal vents have been directly demonstrated in field studies of sociated enzymes have been successfully used as elec- black smoker vents, in situ by harnessing the electrical trodes in manmade fuel cells (e.g., Arechederra and Minteer, current produced by the vent to power a small device on the 2008). As in a PEM fuel cell, the flow of and seafloor (Yamamoto et al., 2013), as well as in the labora- electrons is spatially separated in mitochondria, and life has tory, where it has been demonstrated that black smoker evolved precise chemical means by which to harness the chimney material is electrically conductive and capable of energy present in these potential gradients. The bilayer catalyzing redox reactions (Nakamura et al., 2010b). membrane is the capacitor that separates contrasting che- It is also perhaps pertinent to the PFC model that, as pure mical environments, and the electrodes of mitochondria and metal alloys are frequently found to be highly are the ends of a series of electrochemically connected en- efficient electrocatalysts (Chen et al., 2011), such materials zyme complexes within the inner membrane, together were probably geologically present within the Hadean pe- known as the electron transport chain. The iron-nickel- riod through the reducing power of serpentinizing vents sulfide–containing centers of the enzymes act as and (e.g., , Ni3Fe; Ulrich, 1890; Frost, 1985; Klein (Berg and Holm, 1982; Baymann et al., 2003; and Bach, 2009) and as Fe-Ni meteoritic deposits during the Vignais and Billoud, 2007), while the membranes—capable late heavy bombardment (Buchwald, 1977; Cockell, 2006). as they are of selective ion transport—maintain chemical The corrosion of metals and metal alloys by an electrolyte disequilibrium (Lane and Martin, 2012). Derived from such as seawater can give rise to a localized fuel cell–like proteobacteria, the mitochondria bear certain similarities to system: anodic oxidation of Fe to Fe2 + with concomitant 2- chemiosmotic LUCA (Yang et al., 1985; Martin and Mu¨ller, cathodic reduction of, for example, SO4 to elemental S or 2 - + 1998; de Paula et al., 2013). LUCA’s biological battery S ;H to H2;orNO3 to NO2 (Bryant et al., 2013)—each would have had some similar functional parts, such as the of which could, in principle, establish localized pH/Eh Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. complex I, referred to as the ‘‘steam of the cell’’ by gradients. Efremov and Sazanov (2011), that drives protons to the periplasm through redox disequilibria. These protons form a 3. Creating a Fuel Cell Model of a Prebiotic pH gradient due to the disequilibrium between membrane Hydrogeological System exterior and interior and drive ATP synthase—the rotary biomolecular motor—to make ATP (Yoshida et al., 2001; The energetic ‘‘essence’’ of extant anaerobic life could be Branscomb and Russell, 2013). viewed as quite similar to a fuel cell, in that ion gradients A geological fuel cell phenomenon also can emerge and electron potentials maintained across a membrane pro- without biological mediation at seafloor interfaces, for ex- duce energy as long as fuels and oxidants are supplied. The ample, in hydrothermal vents produced either by magmatic sediment/water or hydrothermal/ocean interfaces in modern THE FUEL CELL MODEL OF ABIOGENESIS 257

marine systems also can constitute naturally occurring fuel cipitate of mixed composition, containing material such as cells, deriving from the of Earth and its , silicates, and iron and other and mediated by microbes acting as redox cat- sulfides and mixed- layered iron ‘‘double’’ oxy- alysts. This has led to the proposal that ion/electron poten- hydroxides (Russell et al., 2014; Fig. 1). Transition metal tials also were the driving forces for life’s origin and that sulfides are one of several possible components of a prebi- these gradients were initially provided by the naturally oc- otic hydrothermal chimney, but as one of the more elec- curring pH/Eh potentials in alkaline hydrothermal vents trochemically active materials likely to be present, it is (Russell and Hall, 1997; Russell et al., 2010). This alkaline useful to characterize iron sulfide chimney growth and hydrothermal origin-of-life model can be generalized to any in the laboratory, which various studies wet rocky world with a chemically similar water-rock in- have pursued (Russell et al., 1989; Mielke et al., 2010, terface (Russell et al., 2014), so we chose to develop our 2011; McGlynn et al., 2012; Barge et al., 2014). Many electrochemical/fuel cell model using alkaline vents as one properties of iron sulfide precipitates relevant to a far-from- example of a hydrogeological origin-of-life scenario that equilibrium geological system have been experimentally could benefit from this concept. Similar fuel cell models determined. For example, when initially precipitated be- also could be applied to simulating electrochemically active tween contrasting solutions, self-assembling iron sulfide water-rock interfaces on early Mars, early Venus, or the icy membranes are capable of generating an electrical potential moons of Jupiter and Saturn. of *0.6 to 0.7 V and maintaining the pH gradient between Earth’s oceans 4 billion years ago are thought to have the hydrothermal and ocean solutions (Filtness et al., 2003; been anoxic, mildly acidic (pH * 5–6 due to dissolved Russell and Hall, 2006; Barge et al., 2014). The precipitates atmospheric CO2, NO, and ephemeral SO2), and rich in produced in out-of-equilibrium experiments have a complex Fe2 + ; and they could have contained Ni2 + ,Mn2 + , phos- structure at the microscale; membranes vary in thickness phates, , nitrite, and some ferric and from *5to*100 lm, exhibit and comprise a mix of iron (Macleod et al., 1994; Russell and Hall, 1997; Hagan et al., sulfides and iron oxyhydroxides/silicates, and can also in- 2007; Martin and Russell 2007; Ducluzeau et al., 2009; corporate dissolved ions such as other transition metals and Mloszewska et al., 2012; Nitschke et al., 2013). The hy- (Mielke et al., 2011; Barge et al., 2012, 2014; drothermal fluid produced by serpentinization would likely McGlynn et al., 2012; Barge and Kanik, unpublished data). have been similar to modern-day hydrothermal fluids: al- In a hydrothermal system with a chimney structure, the kaline, containing some silicate, trace Mo and W, and dis- precipitate could facilitate of ions and electron solved H2 and CH4 along with formate and a range of conduction (Nakamura et al., 2010b); thus, it could function (Proskurowski et al., 2008; Konn et al., 2009; as a selectively permeable electrolyte membrane between

Lang et al., 2010, 2012; Mielke et al., 2010; Charlou et al., hydrothermal fuels and seawater oxidants. In a fuel cell 2010). It is also thought that ancient hydrothermal springs in concept of prebiotic alkaline hydrothermal vents, the anode serpentinizing systems would have contained millimolar and cathode would be the interior and exterior chimney levels of sulfide from dissolution of crustal sulfide , mineral surfaces, respectively. The exterior and interior of a as determined from laboratory reactor experiments simu- chimney could have different mineral compositions, since lating serpentinization in a Hadean system (Mielke et al., chemical garden–like structures tend to exhibit composi- 2010, 2011). Vents driven by serpentinization produce rel- tional variations across the membrane reflecting the chem- atively low-temperature fluids compared to the scalding ical gradients in which they precipitate (Pagano et al., environment of black smokers, but between them, hydro- 2006; Russell and Hall, 2006; Cartwright et al., 2011; Barge thermal vent systems on early Earth could have contributed et al., 2012). Though a hydrothermal chimney would likely many of the fuels and materials that are thought to be rel- also contain components such as silica gel or , the evant for an autogenic/autotrophic emergence of life. The ability of the ‘‘electrodes’’ to drive redox reactions would interfacing of alkaline hydrothermal fluids and seawater on probably be dependent on the presence of electrocatalytic early Earth might have produced a variety of inorganic minerals such as metal sulfides and oxyhydroxides (cf., precipitates within the ambient geochemical/electro- Arrhenius, 2003; Antony et al., 2008). Minerals such as chemical gradients, among them, silica gel, ferrous/ferric (FeS) and greigite (Fe3S4) are electrically oxyhydroxides, and transition metal sulfides (Russell and conductive, can behave as capacitors, and are thought to be Hall, 1997). Hydrothermal precipitates formed by serpenti- catalytically active in organic reactions (Huber and Wa¨ch- nization on early Earth might therefore have had mineral tersha¨user, 1997; Rickard et al., 2001). These iron sulfide compositions and electrochemical properties in some ways minerals also are structurally similar to the Fe4S4 and FeS similar to modern black smoker chimneys, for example, groups in metalloenzymes that catalyze redox reactions in conductive iron sulfide minerals that catalyze redox reac- biology, a similarity that has led to the suggestion that these Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. tions. Thus, electrochemical studies of modern black metalloenzyme centers originally derived from inorganic smokers (e.g., Nakamura et al., 2010b; Yamamoto et al., hydrothermal minerals (Russell and Hall, 1997, 2006; 2013) may also be relevant for understanding the energetics Rothery et al., 2008; Nitschke et al., 2013). In a putative of hydrothermal vent fuel cells on early Earth. prebiotic hydrothermal fuel cell, transport of electrons from A putative Hadean hydrothermal vent with a chimney fuel to oxidant could be mediated by a series of smaller structure would have operated somewhat like a flow-through redox steps, as in biological systems. Within this scenario, fuel cell. The contrasting reservoirs of reduced, H2/CH4- the fuels and oxidants would be indefinitely replenished as containing hydrothermal fluid and the relatively oxidized long as serpentinization remained active, which is envisaged (containing minor concentrations of FeIII and dissolved for modern alkaline hydrothermal systems to last for over NO), CO2-rich ocean would have been separated by a pre- one hundred thousand years (Ludwig et al., 2011). The 258 BARGE ET AL.

FIG. 1. Alkaline hydrothermal vent on early Earth modeled as a membrane fuel cell. The reservoirs of seawater (relatively oxidizing, acidic) and hydrothermal fluid (reducing, alkaline) are physically separated across a hydrothermal chimney pre- cipitate, which functions as part electrolyte and part capacitor. The chimney precipitate would have contained (among other things) iron sulfide minerals, due to the solubility of Fe2 + in the early anoxic oceans and the influx of soluble bisulfide (HS - ) from the hydrothermal fluid. The anode and cathode are the inside and outside surfaces of the chimney, which would have variable composition reflecting the chemical gradients in which they precipitated. The chimney walls would have been somewhat conductive to electrons, which would pass from fuels in the hydrothermal reservoir to oxidants in the ocean reservoir,

catalyzed by electrochemically active minerals in the chimney wall.

chimney membranes themselves could be continuously re- preliminary indications of possible value in the model we newed as the hydrothermal fluids produced by serpentini- propose. zation continue to rise and interface with the ocean, feeding through and possibly disaggregating older membranes. 4.1. Electrochemical studies of simulated hydrothermal precipitates 4. Experimental Studies: Exploiting Electrochemical Though an ancient hydrothermal chimney would have Techniques and Fuel Cells as had a heterogeneous composition, some of the most elec- Test Beds trochemically relevant minerals thought to be formed in this The large body of fuel cell work in which enzymes, mi- system are iron and iron/nickel sulfides. To investigate the tochondria, or microbes are used as electrocatalysts is tes- electrochemical properties of metal sulfide minerals that tament to the importance of charge and concentration could be formed at alkaline vents on early Earth, we pre- gradients in biology. Because both hydrogeological envi- cipitated inorganic membranes in out-of-equilibrium solu- ronments (e.g., vents) and biological cells (at least, in their tion interface experiments (Fig. 2; Filtness et al., 2003; free energy generation mechanisms) can be conceptualized Barge et al., 2014). We interfaced contrasting solutions that as fuel cells, it is reasonable to pursue hydrothermal origin- represent the dissolved Fe2 + in the acidic primordial ocean of-life experiments in a setup that preserves these funda- (added as dissolved FeCl24H2O) and dissolved sulfide in mental aspects of pH and electrochemical gradients across a the alkaline hydrothermal fluid released from crustal sulfide membrane. To this end, we conducted preliminary experi- minerals (added as dissolved Na2S9H2O). When interfaced, ments to simulate hydrogeological systems with electro- these solutions self-assembled into mineral precipitates in Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. chemical gradients and using geological materials as the same manner as the injection chemical garden experi- catalysts. We illustrate how these methods can be ments used previously to simulate Hadean hydrothermal used to simulate geochemical free energy interfaces and test chimney precipitates (Mielke et al., 2011; Barge et al., hypotheses regarding the transition to bioenergetics. It is our 2012, 2014). However, we have found that a glass fuel cell intention to broaden these initial studies of simulated sea- setup yields a more structurally stable membrane precipitate floor/hydrothermal interfaces to explore application of that can be used in subsequent electrochemical experiments similar techniques to other hydrogeological environments, (Barge et al., 2013, 2014). We have previously formed such as galvanic corrosion of iron meteorites or water/rock iron sulfide membranes (and chimneys) using various Fe2 + interfaces on other planets, within the framework of a fuel and sulfide concentrations and have found that, although cell reactor. We envisage the results presented here to be membranes can be formed by using geologically realistic THE FUEL CELL MODEL OF ABIOGENESIS 259

FIG. 2. (A) Precipitation of Fe/S membrane between two interfacing solutions in a glass fuel cell apparatus. The reservoirs represent the Hadean ocean and alkaline hydrothermal fluid, and they are separated by a synthetic porous separator that allows ion contact but not solution mixing. (B) Fe/S membrane precipitated on dialysis membrane (precipitate area *4.9 cm2).

millimolar concentrations over 3–4 days, more robust were carefully rinsed with ddH2O; then NaCl or KCl salt membranes form over shorter timescales when using higher solutions of varying concentration were added to both sides (*50 mM) concentrations, and these still preserve the out- of the precipitated membrane. Salt concentrations were 10 · of-equilibrium nature of the precipitates. Therefore, we used greater on the ‘‘formerly alkaline’’ side than on the ‘‘for- higher reactant concentrations in this study to form precip- merly acidic’’ side, and concentrations between 10 mM and itates on shorter timescales that are reasonable for laboratory 1 M were tested for both NaCl and KCl. Electrodes were experimentation and to generate more precipitate material placed several millimeters from either side of the membrane for use as an electrocatalyst. We recognize that, in a natural surface to record the . In previous

system, similar precipitates might take much longer to form studies, this technique has been shown to generate concen- from the dilute Fe2 + - and sulfide-containing fluids. tration-dependent membrane potentials across inorganic precipitate membranes in many different chemical systems 4.1.1. Formation of membranes simulating hydrothermal including sulfides (e.g., of Hg, Sn), although iron sulfide chimney walls. A synthetic porous membrane template was measurements were not reported (Beg et al., 1977, 1978, clamped between two fluid reservoirs in a two-chamber 1979; Sakashita and Sato, 1977; Malik et al., 1980; Ayalon, glass membrane fuel cell apparatus (Adams & Chittenden 1984; Siddiqi and Alvi, 1989; Kushwaha et al., 2001; Beg glassware, with high-temperature clamp stable up to 150C). and Matin, 2002). Most experiments were performed by using dialysis tubing (Fisher, 3500 MCW) as the precipitation template, but 4.1.3. Membrane galvanostatic tests. We conducted conductive cloth (Sigracet, hydrophilic, no micro- galvanostatic tests in which an electrical current was applied porous layer) was also tested as a template material. The across an iron sulfide membrane after it had formed (Honig iron and sulfide solutions were added to the half-cells re- and Hengst, 1970). Half-cells were filled with acidic spectively, and the cell was kept anoxic throughout with FeCl24H2O solution and alkaline Na2S9H2O solution, constant N2 purging and allowed to react for 24 h at room respectively, and a dialysis membrane was clamped between temperature. High-resolution imaging and chemical analysis the two reservoirs. Two electrodes were placed in each cell: were carried out on the iron sulfide membranes by using an one in the bulk solution to apply current and one very close environmental scanning with an at- (2–3 mm) to the membrane to measure membrane potential. tached energy-dispersive X-ray tube. The electrical potential After the membrane had precipitated for 24 h, a current generated across the inorganic membrane as it precipitated of – 0.75 mA/cm2 was applied across the membrane, and the was recorded over the course of experiments by an Agilent resulting membrane potential relative to the rest potential LXI Data Acquisition/Data Logger Switch Unit with elec- was measured. The measurement also was repeated after Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. trodes placed in each cell. pH measurements were recorded both half-cells had been emptied and refilled with 0.1 M with an Excel XL20 pH/conductivity meter (Fisher Scien- NaCl solution. tific) with temperature calibration. The membranes formed were used in subsequent electrochemical experiments, de- 4.1.4. Voltammetry studies. To investigate the ability of scribed below. iron sulfide chimney precipitates to facilitate electron trans- fer from hydrothermal reductants to seawater oxidants in a 4.1.2. Membrane potential tests. We conducted mem- prebiotic hydrothermal system, we applied simulated hy- brane potential tests in which an iron sulfide membrane was drothermal iron sulfide–containing precipitate membrane allowed to precipitate for 24 h, followed by removal of the material as a catalyst on a glassy carbon electrode (GCE). iron and sulfide solutions. The membrane and both half-cells half-cell experiments were conducted in 260 BARGE ET AL.

which the GCE + catalyst was characterized in a Hadean and separated by an ion-exchange membrane, would allow ocean simulant containing oxidants of interest to the origin for testing of different fluid chemistries and dissolved ionic, of life in serpentinizing systems. In this experiment, a Fe/S as well as gaseous, electron donors and acceptors. MFC- precipitate membrane was formed in the fuel cell apparatus type experiments would also lend themselves well to testing shown in Fig. 2, and after 24 h the membrane was removed, other relevant biological components as catalysts, such as rinsed with ddH2O, and dried under N2. The dried precipi- individual enzymes or . A geo-electrochemical tate was removed from the dialysis membrane template and system also could be simulated by a proton exchange attached to a GCE with conductive carbon tape. Potential membrane fuel cell (PEMFC), which could use gaseous from – 1 V was applied (relative to Ag/AgCl, using a Pt reductants/oxidants under higher pressure or temperature, counter electrode) in a solution representing bicarbonate with the simulated geological catalysts embedded in the gas as an electron acceptor in the Hadean ocean (20 mM diffusion layers (GDLs) within the membrane electrode NaHCO3 + 0.6 M NaCl, titrated to pH *5). A control was assembly (MEA). Either way, the modular engineering ar- performed such that no FeS catalyst was applied to the rangement of fuel cell components makes for a convenient working electrode, but ground FeS particles (Sigma) were planetary geology test-bed system in which components, stirred into the simulated ocean while potential was applied. such as the ion-exchange membranes, electrodes and cata- Other controls had no FeS in the system. Experiments also lysts (or electrolyte assembly unit), and the anode/cathode were performed in which carbon cloth was used as the reactant feeds, may be substituted by materials more closely membrane separator in the two-cell precipitation apparatus connected to geological environments. For example, rather described above, so that Fe/S membranes were precipitated than a commercial GDL catalyst layer, one could substitute directly onto the carbon cloth and pieces of the Fe/S-covered simulated hydrogeological precipitates similar to those de- cloth (5 mm · 50 mm) could be used directly as working scribed above. This type of experiment is not limited to electrodes in voltammetry studies. Controls were also per- simulating hydrothermal vent environments; other geological formed with plain carbon cloth (with no Fe/S) as working electrochemical processes of interest could be explored in a electrode. similar experimental setup, for example, as mentioned in Section 1 above, the electrochemical corrosion of meteorites 4.2. Fuel cell reactors simulating prebiotic on early Earth (Bryant et al., 2009). Field samples of catalytic geo-electrochemical systems minerals derived from reduced Fe/Ni in meteorites, which 4.2.1. Fuel cell design rationale. An origin-of-life fuel have been previously proposed to assist with early phospho- cell experiment designed to test a hydrogeological system rus chemistry (Pasek and Lauretta, 2005; Pasek et al., 2013), on early Earth or any wet rocky might involve a could be treated in the same way as hydrothermal precipitates

combination of techniques: synthesizing simulated mineral in the GDL electrocatalyst layers (see below). catalysts in out-of-equilibrium systems (as in Barge et al., We conducted some preliminary proof-of-concept tests, 2012, 2013, 2014; Mielke et al., 2010, 2011); electro- using a PEMFC to simulate the catalytic action of geolog- chemical studies of simulated precipitate materials to char- ical mineral catalysts in a far-from-equilibrium system. A acterize their utility as electrocatalysts; and the use of these sample experimental arrangement for a PEMFC that simu- simulated prebiotic catalysts to create catalytic electrodes lates a geo-electrochemical system is shown in Fig. 3. In our for use in a fuel cell. There are various types of fuel cell preliminary tests, a commercial proton-exchange membrane experiments that could be informative in this regard. Imi- and two GDLs containing electrocatalysts were sandwiched tating the design of microbial fuel cells (MFCs), where the between two composite bipolar plates, and the fuel catalytic electrodes are submerged in two liquid reservoirs cell was fed with humidified (to represent hydrogen Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only.

FIG. 3. Outline of PEMFC subsystem construction. THE FUEL CELL MODEL OF ABIOGENESIS 261

produced by serpentinization) and air at the anode and cathode sides, respectively. We examined the potential for efficient carbon-black-deposited electrodes to be replaced by geological materials, in this case Fe-Ni alloys present within iron meteorites, where taenite and kamacite are the dominant phases. While these materials may at first glance appear incompatible with terrestrial minerals, there is support for them to have contributed a significant compo- nent of the early Earth (Pasek and Lauretta, 2008), and they are closely related to the hydrothermal Fe- Ni mineral, awaruite (Klein and Bach, 2009). Most impor- tantly, these samples represent a proof of concept for fabricating geologically analogous GDLs or electrodes for a planetary geology fuel cell reactor by using field samples rather than synthetic catalysts. Such a procedure could be applied to other field samples of interest, for ex- ample, samples of hydrothermal precipitates, serpentinite, metal sulfides, or other minerals that might be important in a prebiotic system.

4.2.2. Preparation of geological fuel cell catalysts. The Fe-Ni ‘‘geological electrocatalyst’’ was prepared by using shavings from the coarse octahedrite, group IAB-sLL To- luca meteorite that has a mean composition of 91% Fe and 8.1% Ni (Fig. 4). This geological sample was ball-milled to an average particle diameter of ca. 5 microns before being dispersed in , sonicated in an ultrasonic bath for 10 min, mixed with Nafion binder, and then spray- coated onto a GDL. For both electrodes, the amount of cat- alyst required was calculated based on a 1.0 mg/cm2 catalyst 2

loading within an 11.56 cm active area. Both the cathode and anode electrodes were positioned on the faces of the catalyst-coated Nafion membrane and hot-pressed (with a FIG. 4. (A) Scanning electron microscope image (sec- ondary electron) of shavings from the type IAB-sLL Toluca hydraulic press) to form a unit of MEA. The required meteorite with mean composition of 91% Fe and 8.1% Ni. amount of the catalyst was first ultrasonically mixed for a (B) Energy-dispersive mapping spectrum (EDS) on the few seconds with a few drops of deionized water in order to same course-grained meteorite sample with elemental color break the catalyst powder into small pieces and prevent the identification showing a small degree of surface oxidation of catalyst from being deactivated when blending it with the the meteorite shavings. dispersion agent, ethanol. A calculated amount of 50 wt % ethanol based on 15 mL ethanol per 50 mg catalyst was added to the wet catalyst, and the resulting mixture was heating in 1 M at 80C for 1 h, and finally ultrasonically blended for about 3 min. The resulting blend rinsing several times with deionized water. Both the cathode was then mixed with a 5 wt % Nafion solution (Sigma Al- and anode electrodes were positioned on the faces of the 2 drich) based on a loading of 0.1 mg Nafion/cm . The Nafion Nafion membrane and hot-pressed with a hydraulic press at solution is added to create an ionic phase for proton trans- 130C and 50 kg/cm2 for 3 min to form a unit of MEA. port from the anode to cathode. This final mixture was ul- trasonically blended for 15 min to form what is normally 4.2.4. Assembly and operation of the PEMFC. A single known as an ‘‘electrode ink.’’ The electrode ink was then PEMFC was assembled by sandwiching the MEA between manually sprayed onto the surface of the GDL that was two composite graphite bipolar plates (Bac2) with an 11- mounted on a PTFE plate heated to 80C to evaporate the turn and single-pass serpentine flow channel (2 mm · 2 mm, alcoholic components, until the desired 1.0 mg/cm2 catalyst with a rib of width 0.8 mm). The PEMFC was placed be- loading was reached. tween two end plates and connected to an in-house Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. fuel cell test station. The fuel cell was maintained at room 4.2.3. Preparation of membrane electrode assembly. A temperature and fed with hydrogen at the anode side and air mixture of 5 wt % Nafion solution, calculated based on at the cathode side (Fig. 5). The flow rates of both hydro- 0.5 mg Nafion/cm2, and (about 3–5 mL) was gen and air were kept constant, to a stoichiom- sprayed onto the surface of the catalyzed GDL to enhance etry of 2 calculated based on a 500 mA/cm2 current density. the contact between the catalyst layer and the Nafion Dry hydrogen and air were fully humidified by flowing membrane in the MEA being fabricated. The N115 Nafion through bubbler-type humidifiers. The pressures of the inlet membranes were pre-treated before being used in MEA and outlet gases were 1 and 0 barg, respectively. Electro- fabrication by heating at 80C in 2 wt % chemical tests were performed with a Gamry Reference solution for 1 h, rinsing several times with deionized water, 3000 potentiostat/galvanostat. 262 BARGE ET AL.

FIG. 5. (A) PEMFC test rig at Uni- versity of Leeds. (B) Close-up of com- plete MEA fuel cell assembly with the anodic surface in the foreground con- nected to H2 inlet and outlet supply. The cathodic reverse side is supplied with air, within which O2 acts as ultimate electron acceptor.

5. Results near the interface (forming Fe/S precipitates in the bulk solutions as well as on the carbon cloth), and it was not 5.1. Electrochemical characterization of simulated possible to do electrochemical experiments with the mem- hydrothermal precipitates

brane still in situ. We also found that it was necessary to When the contrasting acidic/iron (ocean simulant) and pretreat the carbon cloth by wetting it with ddH2O, rinsing alkaline/sulfide (hydrothermal simulant) solutions were ad- to remove extraneous carbon material that was not contained ded to the two-chamber cell, they interfaced across the porous within the woven fibers, and applying FeCl24H2O and separating barrier and immediately began to produce a self- Na2S9H2O solutions directly to the cloth to form a assembling iron sulfide–containing precipitate film in/on ‘‘starter’’ precipitate in the pores before placing it between the template. The characteristics of this inorganic mem- the two solution reservoirs. In any case, using carbon cloth brane precipitate varied depending on compositions of the caused the FeS precipitate to form on the carbon cloth fibers two solutions and on the nature of the membrane template (Fig. 6B); and because carbon cloth itself is conductive, a material, but commonly a complex landscape of micron- piece of the cloth containing FeS precipitate could be di- scale crystal structures formed (Fig. 6A)—not unlike the rectly used as a working electrode. crystal morphologies observed in the interior membranes of We observed that, even in a system containing only iron chemical gardens formed in injection experiments (Cart- and sulfide, these self-assembling inorganic membranes wright et al., 2002, 2011; Barge et al., 2012). When dialysis were able to mediate chemical/pH gradients for several tubing was used as the template, the membrane precipitate days, as observed by monitoring the pH of the acidic cell. formed within the pores and also on the surface of the dialysis The membranes also generated a diffusion potential of tubing. Membranes precipitated on dialysis tubing proved to *300–600 mV as they formed, depending on experimental be ideal for electrochemical characterization within the conditions, as was observed in a similar experiment by original precipitation apparatus, since the dialysis template Filtness et al. (2003). This potential declined over several is only permeable to ions and solution mixing was com- days as ions diffused through the membrane, eventually pletely prevented. Thus, precipitates only formed on the reaching equilibrium. When the iron and sulfide solutions template separating the two solution reservoirs, not in either were removed after a membrane had precipitated, and the Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. solution. Dialysis templates also yielded membranes that solution reservoirs were refilled with concentrated salt so- were structurally stable enough to enable gentle rinsing and lutions, the iron sulfide–containing membranes still gener- emptying/refilling of solutions, and when removed from the ated a membrane potential of a few tens to a hundred cell, the precipitate was a flaky solid film that could be millivolts. Consistent with previous studies, this membrane easily dried and removed from the dialysis tubing for later potential was dependent on the concentrations of salts on use as a catalyst. Carbon cloth was much more porous than either side of the membrane and is likely due to the ad- the dialysis tubing, and when used as a precipitation tem- sorption of cations and/or anions on the charged inorganic plate, it did not prevent solution mixing as effectively. Thus, membrane surfaces. In galvanostatic experiments in which in the precipitation fuel cell apparatus where carbon cloth current was applied across a precipitated iron sulfide was used as the separator, the two solutions often mixed membrane, the I/V characteristic curves of these Fe/S THE FUEL CELL MODEL OF ABIOGENESIS 263

FIG. 6. Environmental scanning electron microscope (ESEM) images of iron sulfide membrane precipitates on different membrane templates: (A) dialysis tubing and (B) carbon cloth (the cylinders are carbon fibers). ESEM images were obtained with a voltage of 20 kV and a working distance of 10 mm.

membranes showed that the electrical resistance across the *0.4–0.8 mA of both anodic and cathodic current. No membrane changed at an inflection point (Fig. 7), indi- significant current ( > tens of microamps) was observed with cating that the Fe-S precipitate membranes formed in these just the GCE, GCE with carbon tape, or in the ocean sim- out-of-equilibrium systems are somewhat bipolar, that is, ulant containing FeS particles. Similar results, though at partially cation- and anion-selective. This is an impor- higher current ranging to – 10 mA, were observed when tant property of biological and also electrodialysis mem- carbon cloth containing Fe/S precipitate was directly used as branes (Honig and Hengst, 1970; Sakashita et al., 1983; an electrode. As in the dialysis tubing experiments, the Bauer et al., 1988; Strathmann et al., 1993; Aritomi et al., carbon cloth without Fe/S did not facilitate significant cur- 1996). rent at these potentials. The higher current range of Fe/S The ability of simulated prebiotic iron sulfide chimney membranes on carbon cloth, compared to Fe/S membrane precipitates to facilitate electron transfer to a simulated material taken from dialysis tubing, is likely due to the fact Hadean ocean, with applied potentials between – 1 volt that the active surface area is greater on the carbon cloth (relative to Ag/AgCl), is shown in Fig. 8. The simulated working electrodes. hydrothermal FeS membrane catalyst (grown on dialysis template then attached to a GCE) on the electrode facilitated 5.2. PEMFC function using geological electrocatalyst materials derived from iron meteorites Electrochemical analyses of the PEMFC in operational mode revealed a stable open circuit voltage difference across the electrodes peaking close to 0.7 V (Fig. 9, upper panel). In the presence of Toluca electrocatalyst–impregnated GDLs, the current-voltage polarization curves (Fig. 9, lower panel) show distinct behavior confirming that the fuel cell is func- tional; and the current flow increases with decreasing voltage as expected, approaching a load of ca. 13 mA. In the absence of the geological catalyst components, no such polarization behavior was observed.

6. Discussion Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. The iron sulfide membranes that formed at the interface of contrasting solutions in our experiments represent spe- cific mineral components that could have been contained FIG. 7. Current/voltage characteristics of an iron sulfide within larger, carbonate- and silica-containing hydrothermal precipitate membrane in two repeated experiments. The chimney precipitates on early Earth. Iron sulfide precipitates membrane was formed at the interface between simulated in a hydrothermal chimney, formed in pH/ion gradients Hadean ocean and alkaline hydrothermal solutions; then via diffusion and self-assembly, could contribute to the these solutions were removed and replaced with 0.1 M electrocatalytic properties of the chimney if they were suf- NaCl. Voltages are measured relative to a rest potential. ficiently conductive. Mackinawite (FeS) is usually the first 264 BARGE ET AL.

FIG. 8. Cyclic voltammograms of GCEs in N2-purged Hadean ocean simulant (20 mM NaHCO3 + 0.6 M NaCl, pH 5), with and without simulated Fe/S hydrothermal membrane as an electrocatalyst. Inset (A): anoxic voltammetry cell. Inset (B): Fe/S membrane precipitate dried and applied to GCE.

metastable product that forms from the precipitation of through the membrane. The precipitates formed would aqueous Fe2 + and HS - , and over time it can transform to likely exhibit a compositional gradient from exterior to greigite (Fe3S4), pyrrhotite, or (Kwon et al., 2011; interior, as is observed in real hydrothermal chimneys Csa´kbere´nyi-Malasics et al., 2012; Sines et al., 2012). and chemical garden lab experiments (Haymon, 1983; Mackinawite is a layered mineral composed of Fe-S tetra- Cartwright et al., 2002; Barge et al., 2012), and the charged hedra stabilized by van der Waals forces, and since the mineral surfaces could adsorb and concentrate ions from Fe-Fe distance is only 2.60 A˚ (similar to elemental a-iron: either ocean or hydrothermal solution. Metal sulfide– 2.48 A˚ ), it is highly electrically conductive and has even containing chimneys formed within a temperature/pH gradi- been shown to exhibit superconducting characteristics ent could potentially host a range of mineral phases with (Kwon et al., 2011). It has been demonstrated that black various electrochemical properties. It is likely that these in- smoker chimney material, which is also mostly composed of organic membranes could also incorporate silicates, phos- metal sulfides, is capable of converting the geochemical phates, Ni, and Mo into the precipitate (Barge et al., 2014; redox potential between hydrothermal fluid and seawater to Barge and Kanik, unpublished data). More experimental work electrical current and catalyzing redox reactions (Nakamura is needed to fully characterize the properties of geological et al., 2010b). Even though a hydrothermal chimney at an mineral precipitates in far-from-equilibrium systems in order alkaline vent on early Earth may not have had such a high to understand the reactions that may be possible in hydro- fraction of metal sulfides as a modern black smoker chim- thermal chimney environments. ney, still it is possible that in a prebiotic system subject to In some ways, the inorganic Fe/S precipitates that formed electrical potential gradients between hydrothermal fuels on carbon cloth in these interface experiments are similar to (e.g., H2,CH4) and seawater oxidants (e.g., CO2,NO3 , the ion-exchange membranes and gas diffusion layers in III NO2 ,Fe ), deposits of conductive Fe/S precipitates, aided commercial as well as experimental fuel cells (Kim et al., by their large active surface area as seen in Fig. 6, might 2009; Paulo and Tavares, 2011). The successful use of Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. have been able to facilitate electron conduction and catalyze commercially available fuel cell electrode/GDL material redox reactions on the mineral surfaces. (carbon cloth) as a template to precipitate simulated hy- Our preliminary experiments revealed that, due to their drogeological mineral catalysts is promising, as these fab- far-from-equilibrium nature, inorganic precipitate mem- ricated electrodes can be directly used in PEMFC branes analogous to hydrothermal chimney walls have many assemblies to catalyze redox reactions in a pressurized out- unusual properties (van Oss, 1984) that may be biologically of-equilibrium geology test-bed reactor. Our preliminary and/or prebiotically relevant. The growth of a hydrothermal PEMFC tests showed that geological minerals applied to the chimney precipitate that separates contrasting solutions GDLs as catalysts have the potential to facilitate current could generate and maintain an electrical potential for long flow and possibly catalyze redox reactions in putative geo- periods of time yet still facilitate selective ion transfer logical fuel cells. Our methods proved successful for THE FUEL CELL MODEL OF ABIOGENESIS 265

barriers from CH4 to methanol and CO2 to CO could be solved through electron bifurcation by hydrothermally produced Mo precipitating in trace amounts within a chimney wall (Nitschke and Russell, 2009, 2011; Helz et al., 2011). This is an interesting proposition, since, in biology, energy barriers are also surmounted through elec- tron shuttles such as quinones, flavins, and molybdenum and tungsten pterins (Darrouzet and Daldal, 2002; Staniek et al., 2002; Herrmann et al., 2008; Li et al., 2008; Nitschke and Russell, 2009, 2011, 2013; Kaster et al., 2011; Buckel and Thauer, 2012; Poehlein et al., 2012; Schuchmann and Mu¨ller, 2012). The end result of a fuel cell experiment simulating a geo-electrochemical system would be to reveal redox reactions in either half-cell as a function of the composition of electrodes and membrane. In a simulation of a prebiotic alkaline hydrothermal vent, for example, elec- trical current could be produced by oxidation of hydro- thermal H2 to electrons and protons, thus driving the reduction of oceanic CO2 to formate and possibly further to . Alternatively, hydrothermal CH4 could be oxidized to methanol and formaldehyde perhaps by using nitrate or nitrite as an electron acceptor (Ducluzeau et al., 2009; Nitschke and Russell, 2013; Russell et al., 2013). Organic redox cofactors such as quinones, flavins, NADP, FAD, GMP, and AMP that participate in electron transfer in modern bioenergetics and were likely present in LUCA (Buckel and Thauer, 2012; Schoepp-Cothenet et al., 2013) could also be incorporated into fuel cell electrodes for origin-of-life simulations. Regardless of the specific parameters, approaching origin-

of-life experiments within a fuel cell/electrochemical frame- work could allow experimenters to take advantage of the many rigorous analytical methods and diagnostics that are routinely employed in fuel cell tests. Moreover, there are various computational models that can be used to model fuel cell systems, none of which (to the best of our knowledge) FIG. 9. Open circuit voltage (upper panel) and polari- have been employed with an abiogenesis perspective in zation (I/V) curves of the Toluca MEA fuel cell (lower mind. There are many reactions of importance to the origin panel). The negative sign in the I/V graph indicates that the current is being taken from the functional fuel cell. of life aside from the proto-metabolic redox reactions dis- cussed above, such as RNA synthesis, formation of peptides or other complex organics, and polymerization of phos- phates to create energy currency. Any of these could be creating PEMFC MEAs with geological materials (Toluca incorporated into a fuel cell–like experiment and could meteorite taenite and kamacite material) as electrode cata- benefit from electrochemical techniques. For example, it has lysts and should work for any field sample or synthetic been recently shown that Fe2 + can cause some biological material of similar composition. to become redox active (Athavale et al., 2012; Hsiao Combined, these electrochemical techniques and fuel cell et al., 2013). In an early Earth hydrothermal system, it is reactor designs may allow for testing of specific reactions possible that these redox-active RNAs might participate in that are proposed to be relevant to the emergence of me- the electrocatalytic function of a hydrothermal chimney tabolism. For example, some metal sulfide minerals that system along with the inorganic components (cf. McGlynn form under hydrothermal conditions are structurally similar et al., 2012). Another example could be studying the gen- to the inorganic active centers of certain fundamental redox eration of condensed and/or activated phosphates, which Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. enzymes (Russell and Hall, 1997; Rothery et al., 2008; are necessary at the origin of life to function as energy cur- Baradaran et al., 2013; Nitschke et al., 2013), so synthetic rency molecules similar to ATP (Baltscheffsky, 1996). It has hydrothermal precipitates formed in the laboratory could be been proposed that green , a redox-active iron oxyhy- evaluated as ‘‘proto-metalloenzymes’’ in electrochemical droxide mineral that efficiently absorbs phosphate (Antony and fuel cell studies. Experimentally testing the effects of et al., 2008; Barthe´le´my et al., 2012), might be capable of trace chemical components on the electron transfer capa- forming pyrophosphate bonds via regulation of proton flow bilities of a hydrothermal chimney could also be instruc- through mineral layers (Arrhenius, 2003; Russell et al., tive. To give one example of this, Nitschke and Russell 2013); alternately, pyrophosphate could be formed via sub- (2013) proposed a denitrifying methanotrophic acetogenic strate within a hydrothermally precipitated pathway for the origin of metabolism, where the energy membrane (Martin and Russell, 2007; Barge et al., 2014). 266 BARGE ET AL.

Incorporating components such as these into electrodes in a Author Disclosure Statement fuel cell experiment might yield new prebiotically plausible No competing interests exist. ways to synthesize necessary high-energy molecules.

Abbreviations 7. Conclusions ATP, adenosine triphosphate; GCE, glassy carbon elec- There are many proven engineering design techniques and trode; GDLs, gas diffusion layers; LUCA, last universal analytical methods for building membrane fuel cells and common ancestor; MEA, membrane electrode assembly; characterizing the electrochemical properties of different or- MFCs, microbial fuel cells; PEM, proton exchange mem- ganic and inorganic catalysts. These techniques and methods brane; PEMFC, proton exchange membrane fuel cell; PFC, can offer much insight into the workings of bioenergetics and prebiotic fuel cell. metabolism if applied to biological and geological early Earth . Considering that (1) microbes and mitochondria may be viewed as highly evolved biological fuel cells and (2) References that the earliest life on Earth was likely chemiosmotic, che- Antony, H., Legrand, L., and Chausse´, A. (2008) Carbonate and mosynthetic, and utilized energy sources such as those sulphate green —mechanisms of oxidation and reduc- commonly found in modern alkaline hydrothermal vents, we tion. Electrochim Acta 53:7146–7156. believe that the PFC modeling approach has potential. We Arechederra, R. and Minteer, S.D. (2008) -based argue that the electrochemical methods that have been suc- biofuel cells: immobilized mitochondria on carbon paper cessful in the fields of fuel cell development and bioelec- electrodes. Electrochim Acta 53:6698–6703. trochemistry should be applied to simulating the origin of life Aritomi, T., van den Boomgaard, Th., and Strathmann, H. because of the conspicuous fuel cell–like properties of met- (1996) Current-voltage curve of a bipolar membrane at high abolic and geo-electrochemical systems. The preliminary current density. Desalination 104:13–18. experiments presented here were successful in forming pu- Arndt, N. and Nisbet, E. (2012) Processes on the young Earth tative electrocatalytic minerals in out-of-equilibrium chemi- and the of early life. Annu Rev Earth Planet Sci cal systems, mimicking the formation of metal sulfide 40:521–549. precipitates at the interface between ocean and hydrothermal Arnold, M.A. and Rechnitz, G.A. (1980) Comparison of bac- fluids on early Earth. These inorganic membranes that sim- terial, mitochondrial, , and biocatalysts for ulate prebiotic submarine hydrothermal precipitates are glutamine selective membrane electrodes. Anal Chem 52: shown to be structurally complex, capable of generating and 1170–1174. Arrhenius, G. (2003) Crystals and life. Helv Chim Acta 86:

maintaining membrane potentials and pH gradients, and 1569–1586. electrically conductive. We also successfully formed simu- Athavale, S.S., Petrov, A.S., Hsaio, C., Watkins, D., Prickett, lated hydrothermal chimney precipitates on commercial fuel C.D., Gossett, J.J., Lie, L., Bowman, J.C., O’Neill, E., Ber- cell electrode/GDL material, a technique that can be used to nier, C.R., Hud, N.V., Wartell, R.M., Harey, S.C., and Wil- create electrode components for various types of fuel cell liams, L.D. (2012) RNA folding and catalysis mediated by experiments. Finally, we developed a PEMFC for simulating iron(II). PLoS One 7:e38024. geo-electrochemical processes and demonstrated that the Ayalon, A. (1984) Precipitation membranes—a review. PEMFC is operational when field samples of geological J Membrane Science 20:93–102. material are used as the electrode catalysts. The methods used Baltscheffsky, H. (1996) Energy conversion leading to the or- in this work are easily generalized to other geochemical in- igin and early of life: did inorganic pyrophosphate terfaces of interest, such as water-rock reactions on early precede adenosine triphosphate? In Origin and Evolution of Mars or the icy moons. Moving toward this type of laboratory Biological Energy Conversion, edited by H. Baltscheffsky, simulation of the emergence of bioenergetics will not only be VCH Publishers, Cambridge, pp 1–9. informative in the context of the origin of life on Earth but Baradaran, R., Berrisford, J.M., Minhas, G.S., and Sazanov, may help in understanding whether it is possible for life to L.A. (2013) Crystal structure of the entire respiratory com- have emerged in similar environments on other planets. plex I. Nature 494:443–450. Barbir, F. (2005) PEM Fuel Cells: Theory and Practice, 1st ed., Elsevier Academic Press, Amsterdam. Acknowledgments Barge, L.M., Doloboff, I.J., White, L.M., Russell, M.J., This research was carried out at the Jet Propulsion Lab- Stucky, G.D., and Kanik, I. (2012) Characterization of iron- Langmuir oratory, California Institute of Technology, under a contract phosphate-silicate chemical garden structures. 28:3714–3721. with the National Aeronautics and Space Administration with Barge, L.M., Russell, M.J., and Kanik, I. (2013) Fuel cell Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. support by the NASA Astrobiology Institute (Icy Worlds) and simulations of hydrothermal vents on Europa [abstract 2200]. at the University of Leeds with support from the Centre for In 44th Lunar and Conference Abstracts, Computational Fluid Dynamics and Energy Leeds. We ac- Lunar and Planetary Institute, Houston. knowledge useful discussions with members of the NAI- Barge, L.M., Doloboff, I.J., Russell, M.J., VanderVelde, D., sponsored Disequilibrium and Evolution White, L.M., Kidd, R., and Kanik, I. (2014) Pyrophosphate Focus Group, and we thank two anonymous referees for their synthesis in iron mineral films and membranes simulat- helpful suggestions. J.M.P.H. is a Nuffield Foundation Bursar ing prebiotic submarine hydrothermal systems. Geochim (2013), and L.M.B. is supported by the NAI through the Cosmochim Acta 128:1–12. NASA Postdoctoral Program, administered by Oak Ridge Baross, J.A. and Hoffman, S.E. (1985) Submarine hydrothermal Associated Universities through a contract with NASA. vents and associated gradient environments as sites for the THE FUEL CELL MODEL OF ABIOGENESIS 267

origin and evolution of life. Orig Life Evol Biosph 15: Cartwright, J.H.E., Garcı´a-Ruiz, J.M., Novella, M.L., and 327–345. Ota´lora, F. (2002) Formation of chemical gardens. J Barthe´le´my, K., Naille S., Despas, C., , C., and Mallet, M. Interface Sci 256:351–359. (2012) Carbonated ferric as a new material for Cartwright, J.H.E., Escribano, B., and Sainz-Dı´az, I. (2011) efficient phosphate removal. J Colloid Interface Sci 384: Chemical-garden formation, , and composition. I. 121–127. Effect of the nature of the cations. Langmuir 27:3286–3293. Bauer, B., Gerner, F.J., and Strathmann, H. (1988) Develop- Chang, I.S., Moon, H., Bretschger, O., Jang, J.K., Park, H.I., ment of bipolar membranes. Desalination 68:279–292. Nealson, K.H., and Kim, B.H. (2006). Electrochemically Baymann, F., Lebrun, E., Brugna, M., Schoepp-Cothenet, B., active (EAB) and mediator-less microbial fuel cells. Giudici-Orticoni, M-T., and Nitschke, W. (2003) The redox J Microbiol Biotechnol 16:163–177. construction kit: pre–last universal common ancestor Charlou, J.L., Donval, J.P., Konn, C., Ondre´as, H., and Fouquet, Y. evolution of energy-conserving enzymes. Philos Trans R Soc (2010) High production and fluxes of H2 and CH4 and evidence Lond B Biol Sci 358:267–274. of abiotic synthesis by serpentinization in ultra- Beg, M.N. and Matin, M.A. (2002) Studies with nickel phos- mafic-hosted hydrothermal systems on the Mid-Atlantic Ridge. phate membranes: evaluation of charge density and test of In Diversity of Hydrothermal Systems on Slow Spreading Ocean recently developed theory of membrane potential. J Memb Ridges, Geophysical Monograph Series Vol. 188, edited by P.A. Sci 196:95–102. Rona, C.W. Devey, J. Dyment, and B.J. Murton, American Beg, M.N., Siddiqi, F.A., and Shyam, R. (1977) Studies with Geophysical Union, Washington, DC, pp 265–296. inorganic precipitate membranes: part XIV. Evaluation of Chen, Z., Higgins, D., Yu, A., Zhang, L., and Zhang, J. (2011) effective fixed charge densities. Can J Chem 55:1680–1686. A review on non-precious metal electrocatalysts for PEM fuel Beg, M.N., Siddiqi, F.A., Shyam, R., and Altaf, I. (1978) Stu- cells. Energy Environ Sci 4:3167–3192. dies with inorganic precipitative membranes: XI. Membrane Cockell, C.S. (2006) The origin and emergence of life under potential response, characterization and evaluation of effec- impact bombardment. Philos Trans R Soc Lond B Biol Sci tive fixed charge density. J Electroanal Chem Interfacial 361:1845–1856. Electrochem 89:141–147. Coveney, R.M., Geobel, E.D., Zeller, E.J., Dreschhoff, G.A.M., Beg, M.N., Siddiqi, F.A., Singh, S.P., Prakash, P., and Gupta, and Angino, E.E. (1987) Serpentinization and the origin of V. (1979) Studies with inorganic precipitate membrane: hydrogen gas in Kansas. Am Assoc Pet Geol Bull 71:39–48. evolution of thermodynamically effective fixed charge den- Csa´kbere´nyi-Malasics, D., Rodriguez-Blanco, J.D., Kis, V.K., sity and test of the most recently developed theory of mem- Recnik, A., Benning, L.G. and Po´sfai, M. (2012) Structural brane potential based on the principles of non-equilibrium properties and transformations of precipitated FeS. Chem thermodynamics. Electrochim Acta 24:85–88. Geol 294:249–258. Berg, J.M. and Holm, R.H. (1982) Structure and reactions of Darrouzet, E. and Daldal, F. (2002) Movement of the iron- iron- protein clusters and their synthetic analogs. In sulfur subunit beyond the efLoop of b is required Iron-Sulfur Proteins, edited by T.G. Spiro, John Wiley, New for multiple turnovers of the bc1 complex but not for single York, pp 1–66. turnover Qo site catalysis. J Biol Chem 277:3471–3476. Berg, I.A., Kockelkorn, D., Ramos-Vera, W.H., Say, R.F., de Paula, W.B., Lucas, C.H., Agip, A.N.A., Vizcay-Barrena, G., Zarzycki, J., Hu¨gler, M., Alber, B.E., and Fuchs, G. (2010) and Allen, J.F. (2013) Energy, ageing, fidelity and sex: oocyte Autotrophic carbon fixation in . Nat Rev Microbiol mitochondrial DNA as a protected genetic template. Philos 8:447–460. Trans R Soc Lond B Biol Sci 368, doi:10.1098/ Bradley, A.S. and Summons, R.E. (2010) Multiple origins of rstb.2012.0263. at the Lost City Hydrothermal Field. Earth Planet Ducluzeau, A.-L., van Lis, R., Duval, S., Schoepp-Cothenet, B., Sci Lett 297:34–41. Russell, M.J., and Nitschke, W. (2009) Was nitric the Branscomb, E. and Russell, M.J. (2013) Turnstiles and bi- first strongly oxidizing terminal electron sink? Trends furcators: the disequilibrium converting that put Biochem Sci 34:9–15. metabolism on the road. Biochim Biophys Acta Bioenergetics Eck, R.V. and Dayhoff, M.O. (1966) Evolution of the structure 1827:62–78. of based on living relics of primitive amino acid Bryant, D.E., Greenfield, D., Walshaw, R.D., Evans, S.M., sequences. Science 152:363–366. Nimmo, A.E., Smith, C.L., Wang, L., Pasek, M.A., and Kee, Efremov, R.G. and Sazanov, L.A. (2011) Respiratory T.P. (2009) Electrochemical studies of iron meteorites: complex I: ‘steam engine’ of the cell? Curr Opin Struct Biol redox chemistry on the early Earth. International 21:532–540. Journal of Astrobiology 8:27–36. Ehlmann, B.L., Mustard, J.F., and Murchie, S.L. (2010) Geo- Bryant, D.E., Greenfield, D., Walshaw, R.D., Johnson, B.R.G., logic setting of serpentine deposits on Mars. Geophys Res Herschy, B., Smith, C., Pasek, M.A., Telford, R., Scowen, I., Lett 37, doi:10.1029/2010GL042596. Munshi, T., Edwards, H.G.M., Cousins, C.R., Crawford, I.A., El-Naggar, M.Y., Wanger, G., Leung, K.M., Yuzvinsky, T.D.,

Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. and Kee, T.P. (2013) Hydrothermal modification of the Southam, G., Yang, J., Lau, W.M., Nealson, K.H., and Gorby, Sikhote-Alin iron meteorite under low pH geothermal envi- Y.A. (2010) Electrical transport along bacterial nanowires ronments. A plausibly prebiotic route to activated phosphorus from Shewanella oneidensis MR-1. Proc Natl Acad Sci USA on the early Earth. Geochim Cosmochim Acta 109:90–112. 107:18127–18131. Buchwald, V.F. (1977) The of iron meteorites. Etiope, G. and Sherwood-Lollar, B. (2013) Abiotic methane on Philos Transact A Math Phys Eng Sci 286:453–491. Earth. Rev Geophys 51:276–299. Buckel, W. and Thauer, R.K. (2012) Energy conservation via Etiope, G., Vance, S., Christensen, L.E., Marques, J.M., and electron bifurcating ferredoxin reduction and proton/Na + Ribeiro da Costa, I. (2013) Methane in serpentinized ultra- translocating ferredoxin oxidation. Biochim Biophys Acta mafic rocks in mainland Portugal. Marine and Bioenergetics 1827:94–113. Geology 45:12–16. 268 BARGE ET AL.

Filtness, M.J., Butler, I.B., and Rickard, D. (2003) The origin of A.S., Brazelton, W.J., Roe, K., Elend, M.J., Delacour, A., life: the properties of iron sulphide membranes. Applied Bernasconi, S.M., Lilley, M.D., Baross, J.A., Summons, R.E., Earth Science (Trans Inst Min Metall B) 112:B171–B172. and Sylva, S.P. (2005) A serpentinite-hosted : the Frost, B.R. (1985) On the stability of sulfides, , and na- Lost City Hydrothermal Field. Science 307:1428–1434. tive metals in serpentinite. Journal of 26:31–63. Kim, T.W., Sahimi, M., and Tsotsis, T.T. (2009) Preparation Fuchs, G. (1989) Alternative pathways of autotrophic CO2 and characterization of hybrid -sulfonated poly- fixation. In Autotrophic Bacteria, edited by H.G. Schlegel etheretherketone (SPEEK) cation-exchange membranes. Ind and B. Bowen, Science Technology, Madison, WI, pp 365– Eng Chem Res 48:9504–9513. 382. Klein, F. and Bach, W. (2009) Fe-Ni-Co-O-S phase relations Fuchs, G. (2011) Alternative pathways of in -seawater interactions. Journal of Petrology 50: fixation: insights into the early evolution of life. Annu Rev 37–59. Microbiol 65:631–658. Klein, F., Bach, W., and McCollom, T.M. (2013) Composi- Gogarten, J.P., Kibak, H., Dittrich, P., Taiz, L., Bowman, E.J., tional controls on hydrogen generation during serpentiniza- Bowman, B.J., Manolson, M.F. Poole, R.J., Date, T., and tion of ultramafic rocks. Lithos 178:55–69. Oshima, T. (1989) Evolution of the vacuolar H + -ATPase: Konn, C., Charlou, J.L., Donval, J.P., Holm, N.G., Dehairs, F., implications for the origin of eukaryotes. Proc Natl Acad Sci and Bouillon, S. (2009) Hydrocarbons and oxidized organic USA 86:6661–6665. compounds in hydrothermal fluids from Rainbow and Lost Hagan, W.J., Parker, A., Steuerwald, A., and Hathaway, M. City ultramafic-hosted vents. Chem Geol 258:299–314. (2007) Phosphate solubility and the cyanate-mediated syn- Kushwaha, R.S., Ansari, M.A., and Tiwari, R. (2001) Transport thesis of pyrophosphate. Orig Life Evol Biosph 37:113–122. studies of parchment supported model membrane: test of bi- Haile, S.M. (2003a) Fuel cell materials and components. Acta ionic potential theories and evaluation of membrane selec- Mater 51:5981–6000. tivity for metal ions. Indian J Chem 40A:270–274. Haile, S.M. (2003b) Materials for fuel cells. Mater Today Kwon, K.D., Refson, K., Bone, S., Qiao, R., Yang, W.L., Liu, 18:24–29. Z., and Sposito, G. (2011) Magnetic ordering in tetragonal Harold, F. (1986) The Vital Force, Freeman, New York. FeS: evidence for strong itinerant spin fluctuations. Phys Rev Haymon, R.M. (1983) Growth history of hydrothermal black B Condens Matter Mater Phys 83, doi:10.1103/Phys- smoker chimneys. Nature 301:695–698. RevB.83.064402. Heller, A. (1992) Electrical connection of enzyme redox centers Lane, N. and Martin, W.F. (2012) The origin of membrane to electrodes. J Phys Chem 96:3579–3587. bioenergetics. Cell 151:1406–1416. Helz, G.R., Bura-Nakic´, E., Mikac, N., and Ciglenecki, I. Lang, S.Q., Butterfield, D.A., Schulte, M., Kelley, D.S., and (2011) New model for molybdenum behavior in euxinic Lilley, M.D. (2010) Elevated concentrations of formate, ac- . Chem Geol 284:323–332. etate and dissolved organic carbon found at the Lost Herrmann, G., Jayamani, E., Mai, G., and Buckel, W. (2008) City hydrothermal field. Geochim Cosmochim Acta 74:941– Energy conservation via electron-transferring flavoprotein in 952. anaerobic bacteria. J Bacteriol 190:784–791. Lang, S.Q., Fru¨h-Green, G.L., Bernasconi, S.M., Lilley, M.D., Honig, E.P. and Hengst, J.H. (1970) Current/voltage curves of Proskurowski, G., Me´hay, S., and Butterfield, D.A. (2012) BaSO4 membranes. Electrochim Acta 15:491–499. Microbial utilization of abiogenic carbon and hydrogen in Hsiao, C., Chou, I.C., Okafor, C.D., Bowman, J.C., O’Neill, a serpentinite-hosted system. Geochim Cosmochim Acta E.B., Athavale, S.S., Petrov, A.S., Hud, N.V., Wartell, R.M., 92:82–99. Harvey, S.C., and Williams, L.D. (2013) RNA with iron(II) Li, F., Hinderberger, J., Seedorf, H., Zhang, J., Buckel, W., and as a catalyses electron transfer. Nat Chem 5:525– Thauer, R.K. (2008) Coupled ferredoxin and crotonyl coen- 528. zyme A (CoA) reduction with NADH catalyzed by the butyryl- Huang, Y., He, Z., Kan, J., Manohar, A.K., Nealson, K.H., and CoA dehydrogenase/Etf complex from Clostridium kluyveri. Mansfeld, F. (2012) Electricity generation from a floating J Bacteriol 190:843–850. . Bioresour Technol 114:308–313. Lill, S.O.N. and Siegbahn, P.E.M (2009) An autocatalytic Huber, C. and Wa¨chtersha¨user, G. (1997). Activated mechanism for NiFe-: reduction to Ni(I) fol- by carbon fixation on (Fe,Ni)S under primordial conditions. lowed by . 48:1056–1066. Science 276:245–247. Ludwig, K.A., Kelley, D.S., Butterfield, D.A., Nelson, B.K., Jagendorf, A.T. and Uribe, E. (1966) ATP formation caused by and Fru¨h-Green, G. (2006) Formation and evolution of acid-base transition of spinach . Proc Natl Acad carbonate chimneys at the Lost City Hydrothermal Field. Sci USA 55:170–177. Geochim Cosmochim Acta 70:3625–3645. Kaster, A.-K., Moll, J., Parey, K., and Thauer, R.K. (2011) Ludwig, K.A., Shen, C.-C., Kelley, D.S., Cheng, H., and Ed- Coupling of ferredoxin and heterodisulfide reduction via wards, R.L. (2011) U–Th and 230Th ages of electron bifurcation in hydrogenotrophic methanogenic carbonate chimneys at the Lost City Hydrothermal Field.

Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. Archaea. Proc Natl Acad Sci USA 108:2981–2986. Geochim Cosmochim Acta 75:1869–1888. Kelley, D.S., Karson J.A., Blackman, D.K., Fru¨h-Green, G.L., Macleod, G., McKeown, C., Hall, A.J., and Russell, M.J. (1994) Butterfield, D.A., Lilley, M.D., Olson, E.J., Schrenk, M.O., Hydrothermal and oceanic pH conditions of possible rele- Roe, K.K., Lebon, G.T., Rivizzingo, P., and the AT3-60 vance to the origin of life. Orig Life Evol Biosph 24:19–41. Shipboard Party. (2001) An off-axis hydrothermal vent field Malik, W.U., Srivastava, S.K., Bhandari, V.M., and Kumar, S. near the Mid-Atlantic Ridge at 30N. Nature 412:145–149. (1980) Electrochemical properties of zirconyl tungstate Kelley, D.S., Karson, J.A., Fru¨h-Green, G.L., Yoerger, D.R., membrane. J Electroanal Chem Interfacial Electrochem Shank, T.M., Butterfield, D.A., Hayes, J.M., Schrenk, M.O., 110:181–203. Olson, E.J., Proskurowski, G., Jakuba, M., Bradley, A., Lar- Martin, W. and Mu¨ller, M. (1998) The hydrogen hypothesis for son, B., Ludwig, K., Glickson, D., Buckman, K., Bradley, the first . Nature 392:37–41. THE FUEL CELL MODEL OF ABIOGENESIS 269

Martin, W. and Russell, M.J. (2007) On the origin of bio- substrates in bioenergetics. Biochim Biophys Acta Bioener- chemistry at an alkaline hydrothermal vent. Philos Trans R getics 1827:871–881. Soc Lond B Biol Sci 362:1887–1925. Pagano, J.J., Thouvenel-Romans, S.T., and Steinbock, O. Martin, W., Baross, J., Kelley, D., and Russell, M.J. (2008) (2006) Compositional analysis of copper-silica precipitation Hydrothermal vents and the origin of life. Nat Rev Microbiol tubes. Phys Chem Chem Phys 9:110–116. 6:806–814. Pasek, M.A. and Lauretta, D.S. (2005) Aqueous corrosion of McGlynn, S.E., Mulder, D.W., Shepard, E.M., Broderick, J.B., phosphide minerals from iron meteorites: a highly reactive and Peters, J.W. (2009) Hydrogenase cluster : source of prebiotic phosphorus on the surface of the early organometallic chemistry nature’s way. Dalton Trans Earth. Astrobiology 5:515–535. 2009:4274–4285. Pasek, M.A. and Lauretta, D.S. (2008) Extraterrestrial flux of McGlynn, S.E., Kanik, I., and Russell, M.J. (2012) Modification potentially prebiotic C, N, and P to the early Earth. Orig Life of simulated hydrothermal iron sulfide chimneys by RNA and Evol Biosph 38:5–21. peptides. Philos Transact A Math Phys Eng Sci 370:1–16. Pasek, M.A., Harnmeijer, J.P., Buick, R., Gull, M., and Atlas, Z. Michalski, J.R., Cuadros, J., Niles, P.B., Parnell, J., Rogers, (2013) Evidence for reactive reduced phosphorus in A.D., and Wright, S.P. (2013) Groundwater activity on the early ocean. Proc Natl Acad Sci USA Mars and implications for a deep . Nat Geosci 6: 110:10089–10094. 133–138. Paulo, M.J.V.R. and Tavares, A. (2011) Novel hydrotalcite Mielke, R.E., Russell, M.J., Wilson, P.R., McGlynn, S., Cole- composites membranes for alkaline fuel cells. ECS Trans man, M., Kidd, R., and Kanik, I. (2010) Design, fabrication 35:21–28. and test of a hydrothermal reactor for origin-of-life experi- Poehlein, A., Schmidt, S., Kaster, A.K., Goenrich, M., Voll- ments. Astrobiology 10:799–810. mers, J., Thurmer, A., Bertsch, J., Schuchmann, K., Voigt, B., Mielke, R.E., Robinson, K.J., White, L.M., McGlynn, S.E., Hecker, M., Daniel, R., Thauer, R.K., Gottschalk, G., and McEachern, K., Bhartia, R., Kanik, I., and Russell, M.J. Mu¨ller, V. (2012) An ancient pathway combining carbon (2011) Iron-sulfide-bearing chimneys as potential catalytic dioxide fixation with the generation and utilization of a so- energy traps at life’s emergence. Astrobiology 11:933–950. dium ion gradient for ATP synthesis. PLoS One 7:e33439. Mitchell, P. (1961) Coupling of phosphorylation to electron and Proskurowski, G., Lilley, M.D., Seewald, J.S., Fru¨h-Green, hydrogen transfer by a chemiosmotic type of mechanism. G.L., Olson, E.J., Lupton, J.E., Sylva, S.P., and Kelley, D.S. Nature 191:144–148. (2008) Abiogenic hydrocarbon production at Lost City Mitchell, P. (1977) A commentary on alternative hypotheses of Hydrothermal Field. Science 319:604–607. protonic coupling in the membrane systems catalysing oxida- Rickard, D., Butler, I.B., and Oldroyd, A. (2001) A novel iron tive and photosynthetic phosphorylation. FEBS Lett 78:1–20. sulphide mineral switch and its implications for Earth and Mloszewska, A.M., Pecoits, E., Cates, N.L., Mojzsis, S.J., planetary science. Earth Planet Sci Lett 189:85–91. O’Neil, J., Robbins, L.J., and Konhauser, K.O. (2012) The Rothery, R.A., Workun, G.J., and Weiner, J.H. (2008) The composition of Earth’s oldest iron formations: the Nuvvua- prokaryotic complex iron–sulfur molybdoenzyme family. gittuq Supracrustal Belt (Que´bec, Canada). Earth Planet Sci Biochim Biophys Acta 1778:1897–1929. Lett 317–318:331–342. Russell, M.J. and Hall, A.J. (1997) The emergence of life from Mulkidjanian, A.Y., Makarova, K.S., Galperin, M.Y., and iron monosulphide bubbles at a submarine hydrothermal re- Koonin, E.V. (2007) Inventing the dynamo machine: the dox and pH front. J Geol Soc Lond 154:377–402. evolution of the F-type and V-type ATPases. Nat Rev Russell, M.J. and Hall, A.J. (2006) The onset and early evolu- Microbiol 5:892–899. tion of life. In Evolution of Early Earth’s Atmosphere, Hy- Nakamura, R., Okamoto, A., Tajima, N., Newton, G.J., Kai, F., drosphere, and Biosphere—Constraints from Deposits, Takashima, T., and Hashimoto, K. (2010a) Biological iron- Geological Society of America Memoir 198, edited by S.E. monosulfide production for efficient electricity harvesting Kesler and H. Ohmoto, Geological Society of America, from a deep-sea metal-reducing bacterium. Chembiochem Boulder, CO, pp 1–32. 11:643–645. Russell, M.J., Hall, A.J., and Turner, D. (1989) In vitro growth Nakamura, R., Takashima, T., Kato, S., Takai, K., Yamamoto, of iron sulphide chimneys: possible culture chambers for M., and Hashimoto, K. (2010b) Electrical current generation origin-of-life experiments. Terra Nova 1:238–241. across a black smoker chimney. Angew Chem Int Ed Engl Russell, M.J., Daniel, R.M., Hall, A.J., and Sherringham, J. 49:7692–7694. (1994) A hydrothermally precipitated catalytic iron sulphide Neal, C. and Stanger, G. (1983) Hydrogen generation from membrane as a first step toward life. J Mol Evol 39:231–243. mantle source rocks in Oman. Earth Planet Sci Lett 66:315– Russell, M.J., Hall, A.J., and Martin, W. (2010) Serpentiniza- 320. tion as a source of energy at the origin of life. Nitschke, W. and Russell, M.J. (2009) Hydrothermal focusing 8:355–371. of chemical and chemiosmotic energy, supported by delivery Russell, M.J., Nitschke, W., and Branscomb, E. (2013) The

Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only. of catalytic Fe, Ni, Mo/W, Co, S and Se, forced life to inevitable journey to being. Philos Trans R Soc Lond B Biol emerge. J Mol Evol 69:481–496. Sci 368, doi:10.1098/rstb.2012.0254. Nitschke, W. and Russell, M.J. (2011) Redox bifurcations; how Russell, M.J., Barge, L.M., Bhartia, R., Bocanegra, D., Bracher, they work and what they mean to extant life and (potentially) P.J., Branscomb, E., Kidd, R., McGlynn, S., Meier, D.H., to its inorganic roots. Bioessays 34:106–109. Nitschke, W., Shibuya, T., Vance, S., White, L., and Kanik, I. Nitschke, W. and Russell, M.J. (2013) Beating the acetyl co- (2014) The drive to life on wet and icy worlds. Astrobiology enzyme-A pathway to the origin of life. Philos Trans R Soc 14, doi:10.1089/ast.2013.1110. Lond B Biol Sci 368, doi:10.1098/rstb.2012.0258. Sakashita, M. and Sato, N. (1977) The effect of molybdate Nitschke, W., McGlynn, S.E., Milner-White, J.E, and Russell, anion on the ion-selectivity of hydrous ferric oxide films in M.J. (2013) On the antiquity of metalloenzymes and their chloride solutions. Corros Sci 17:473–486. 270 BARGE ET AL.

Sakashita, M., Fujita, S., and Sato, N. (1983) Current-voltage Vignais, P.M. and Billoud, B. (2007) Occurrence, classification, characteristics and oxide formation of bipolar PbSO4 pre- and biological function of : an overview. Chem cipitate membranes. J Electroanal Chem Interfacial Elec- Rev 107:4206–4272. trochem 154:273–280. Volbeda, A. and Fontecilla-Camps, J.C. (2006) Catalytic nickel– Say, R.F. and Fuchs, G. (2010) Fructose 1,6-bisphosphate iron–sulfur clusters: from minerals to enzymes. In Topics in aldolase/phosphatase may be an ancestral gluconeogenic Organometallic Chemistry, Vol. 17, edited by G. Simmon- enzyme. Nature 464:1077–1081. neaux, Springer, Berlin, pp 57–82. Schoepp-Cothenet, B., van Lis, R., Atteia, A., Baymann, F., Ca- Williams, R.J.P. (1961) Possible functions of chains of cata- powiez, L., Ducluzeau, A-L., Duval, S., ten Brink, F., Russell, lysts. J Theor Biol 1:1–17. M.J., and Nitschke, W. (2013) On the universal core of bioen- Williams, R.J.P. (1965) Electron migration in iron compounds. ergetics. Biochim Biophys Acta Bioenergetics 1827:79–93. In Non- Iron Proteins: Role in Energy Conversion, Schuchmann, K. and Mu¨ller, V. (2012) A bacterial electron- edited by A. San Pietro, The Antioch Press, Yellow Springs, bifurcating hydrogenase. J Biol Chem 287:31165–31171. OH, pp 7–22. Siddiqi, F.A. and Alvi, N.I. (1989) Transport studies with parch- Wray, J.J. and Ehlmann, B.L. (2011) Geology of possible ment supported model membranes. JMembSci46:185–202. martian methane source regions. Planet Space Sci 59:196– Sines, I.T., Vaughn, D.D., II, Misra, R., Popczun, E.J., and 202. Schaak, R.E. (2012) Synthesis of tetragonal mackinawite- Yamamoto, M., Nakamura, R., Oguri, K., Kawagucci, S., Su- type FeS nanosheets by solvothermal crystallization. J Solid zuki, K., Hashimoto, K., and Takai, K. (2013) Generation of State Chem 196:17–20. electricity and illumination by an environmental fuel cell Sleep, N.H., Bird, D.K., and Pope, E.C. (2011) Serpentinite and in deep-sea hydrothermal vents. Angew Chem Int Ed Engl 52: the dawn of life. Philos Trans R Soc Lond B Biol Sci 1–5. 366:2857–2869. Yang, D., Oyaizu, Y., Oyaizu, H., Olsen, G.J., and Woese, C.R. Staniek, K., Gille, L., Kozlov, A.V., and Nohl, H. (2002) Mi- (1985) Mitochondrial origins. Proc Natl Acad Sci USA tochondrial formation is controlled by 82:4443–4447. electron bifurcation to the high and low potential pathways. Yoshida, M., Muneyuki, E., and Hisabori, T. (2001) ATP Free Radic Res 36:381–387. synthase—a marvellous rotary engine of the cell. Nat Rev Strathmann, H., Rapp, H.-J., Bauer, B., and Bell, C.M. (1993) Mol Cell Biol 2:669–677. Theoretical and practical aspects of preparing bipolar mem- branes. Desalination 90:303–323. Tran, P.D. and Barber, J. (2012) Proton reduction to hydrogen in biological and chemical systems. Phys Chem Chem Phys Address correspondence to: 14:13772–13784. Laura M. Barge Ulrich, G.H.F. (1890) On the discovery, mode of occurrence Mail Stop 183-601 and distribution of nickel-iron awaruite on the west Jet Propulsion Laboratory coast of South Island, New Zealand. Quarterly Journal of the California Institute of Technology Geological Society of London 46:619–633. 4800 Oak Grove Dr. van Oss, C.J. (1984) Specifically impermeable precipitate Pasadena, CA 91109 membranes. In Surface and Colloid Science, Vol. 13, Springer, New York, pp 115–144. E-mail: [email protected] Vance, S., Harnmeijer, J., Kimura, J., Hussmann, H., Demartin, B., and Brown, J.M. (2007) Hydrothermal systems in small Submitted 4 November 2013 ocean planets. Astrobiology 7:987–1005. Accepted 3 February 2014 Downloaded by OKLAHOMA STATE UNIVERSITY from online.liebertpub.com at 09/11/17. For personal use only.