1. Introduction

Many scholars argue that genetic replication alone will not make a realistic scenario for the first Darwinian process and the start of molecular evolution (e.g. Dyson 1986). So "genes first" should be complemented by some kind of metabolism to create the necessary living and replicating molecules; i.e. "metabolism first". But what does "metabolism first" really mean? We want to argue here that "metabolism", although generally not very well defined, is often used rather as a synonym for "good" boundary conditions to host life. However, based on an understanding of the entropic need for disequilibrium boundary conditions, we can arrive at a more precise definition of what is necessary to create information bearing molecules and to drive a Darwinian Process of replication, mutation and selection. The second law of thermodynamics requires disequilibrium settings to sustain life and this has been known for decades (Schrödinger 1946, Wicken 1978). We believe an understanding of "metabolism first"complements "genetic first" by providing the necessary boundary conditions.

Old ideas on the origin of life were dominated by the belief in 'spontaneous generation', i.e. the idea that life transforms itself all the time from non-living matter. Animals were supposed to appear out of all kinds of non-living materials and the idea that life may have arisen slowly, from an ancient earth after a long evolutionary period of development were utterly foreign. The basic idea, which has lived on into the 21st century, is that chemicals may be mixed together, and become energized thereby achieving life, and this belief has been rightly criticized as implausible (Joseph and Schild 2010). And yet, many in the scientific community are still preoccupied with a notion of mixing things together—that life can be created spontaneously if only we can divine the right recipe and discover and mix together the right concentrations of the correct and matching molecules.

These rather naive conceptions were long ago supplanted by more plausible scenarios, including the "RNA World" hypothesis which incorporates conceptions of entropy and energy. But is such a picture of life's origins accurate? Does it create enough structure to drive a process of Darwinian evolution? Our guess is: No. Loss of molecules by decay will prevent life from arising. Furthermore, simple diffusion is a very powerful driving force on the small micrometer scale. Both have to be compensated for by a disequilibrium condition to refresh the molecules. While such mechanisms and considerations are fundamental, they are rarely considered by those concerned with the origin of life.

2. Equilibrium Settings

Vesicles (Figure 1a) have been discussed at length as the be all and end all of the origin of life (Koch 1985, Szathmáry and Demeter, 1987, Szostak 2001). This idea could be called the "Vesicles First" hypothesis, i.e. that compartmentalization and encapsulation enable life to achieve life, once lipids were joined together to create membranes. Although very interesting properties of lipids have been reported (Walde et al. 1994, Oberholzer et al. 1995, Wick and Luisi 1996, Tsukahara et al. 2004), we find its initial importance doubtful. The main reason being the isolation of molecules into an equilibrium state (Figure 1a). No enhanced concentration of molecules is achieved by the presence of a membrane enclosure. If the concentration is not high at the outside it will not flow into the inside and the vesicles cannot be filled to high concentrations since a directional transport is lacking the necessary highly evolved protein machinery and the energy to fire transport. While the feeding of replication or translation from the outside has been demonstrated, these experiments first required a filling of the vesicles at high concentrations (Hanczyc et al. 2003). At these high concentrations, the experiments could have been performed without using a membrane enclosure and therefore mean nothing. With the membranes, molecular reactions are rather hindered, likewise, the replacement of degraded molecules with new ones is inhibited.

Many scientists reject the "Vesicles First" hypothesis and believe that vesicles were not necessary for the development of the first replicating molecules (e.g. Martin and Russell 2003), such that cellular organization came later, not first. Manfred Eigen (1981) has argued that "organization into cells was surely postponed as long as possible. Anything that interposed spatial limits in a homogeneous system would have introduced difficult problems for prebiotic chemistry. Constructing boundaries, transposing things across them and modifying them when necessary are tasks accomplished today by the most refined cellular processes." Vesicles are considered too much “equilibrium” than “non-equilibrium” boundary conditions.

Similarly, the absorption on surfaces (Figure 1b) has been discussed as a mechanism to enhance the local concentration of the necessary chemicals and to trigger catalytic activity (Ferris and Ertem 1992, Kawamura and Ferris 1994). However experiments using longer, information bearing molecules typically require a washing step after surface catalysis to remove the molecules. Although a disequilibrium boundary condition which could implement such a desorption mechanism might provide the necessary impetus for life, this possibility is seldom taken into consideration. Therefore, according to the surface absorption hypothesis, the molecules have to sit on the surface until they are degraded into pieces short enough such that they can be removed by desorption using Brownian motion. However, in fact, the templated replication of longer nucleotides (Figure 1c) poisons the replicated template and they will only be removed by degradation.

The theory of spontaneous generation has never really died, although one might have expected the experiments of Louis Pasteur to have administered the coup de grâce with his work "On the organized corpuscles that exist in the atmosphere; examination of the doctrine of spontaneous generation" (Louis Pasteur 1861). Pasteur was able to show that spores and bacterial life forms exist on the microscale and induce only an apparent "spontaneous" generation of life. We see an analogy here with present day thinking. Just as people at that time took for granted that life creates itself everywhere from scratch, origin-of-life researchers tend to take it for granted that life should be able to grow from scratch, or at least from very limited food sources and in equilibrium settings. These views have been ridiculed as completely implausible (Joseph 2009; Joseph and Schild 2010).

Life can only survive when it has ways to import food into a reaction chamber against a considerable chemical potential created by high concentrations of molecules already inside the cell. This is made possible via directed membrane transport equipped with highly evolved protein machinery. How could a primitive membrane machinery create directed transport across a membrane without an ATP metabolism and in the absence of highly evolved membrane proteins? On the other hand, as we will see below, rather simple disequilibrium conditions such as a thermal gradient, can provide such a pump and import molecules into microscale pores in mineral precipitates (Baaske et al. 2007).

3. Disequilibrium Settings

The examples of Table 1 illustrate the point we want to make here. Without doubt, most origin of life experiments do indeed use disequilibrium conditions in the lab. For example, they start with a high molecule concentration, fill molecules at high concentration into vesicles, put molecules on unloaded, fresh catalytic surfaces and begin after creating well prepared initial conditions. However, these conditions lack a persistent, geological disequilibrium implementation. The experiments run into thermodynamic equilibrium, or in other words, they die. They miss the opportunity to explore continuous evolution in a maintained disequilibrium steady state, and therefore, fail to create the conditions necessary for the establishment of life.

Persistent disequilibrium is for example obtained from a constant gradient of a thermodynamic property. This can be a temperature gradient (as discussed later) or the close contact of two different materials which induce salt, pH and redox potentials. The latter is discussed in the geometry of membranous boundaries of mineral precipitate enclosures are envisioned by Michael Russell (Russell and Hall, 1997, Martin and Russell 2003, Martin and Russell 2007). These chambers have built-in gradients across a semi-permeable membrane. Directed electrical transport with the possibility of catalytic synthesis along the pH gradient is the natural result of a geologically focused gradient. We envision that many other disequilibrium settings are likely to be found, discussed and explored in the future as the advantages of disequilibrium settings for the origin of life are recognized.

Another persistent disequilibrium can occur in a flow from a warm liquid to a cold one, a setting studied for example by Koichiro Matsuno, who demonstrated the polymerization of peptides in the fast cooling environment of hydrothermal outflows into the ocean (Imai et al. 1999). Other examples are dry-wet cycles or day-night cycles which induce oscillatory temperature or phase state changes on the hour time scale. We are not a particular fan of the latter setting as the volume which is used for wetting will lead to very strong dilutions that are not compensated for in the next drying cycle. As a result, the overall loss of molecules is probably too large to be compensated by the gain of concentration in the dry cycle. Also, like day and night cycles, the kinetics might be too slow to yield a fast evolving system.

Shown in Figure 2 is the example of a disequilibrium setting of a temperature gradient. It can suck a high concentration of molecules (Braun and Libchaber 2002, Baaske et al. 2007) into a millimeter-sized chamber and concurrently implement a thermal cycling in the same setting to foster DNA replication (Braun et al. 2003) (Figure 2a). As demonstrated, both can be combined (Mast and Braun, 2010) to replicate DNA by thermal cycles in an exponential manner and the accumulation of the replicates in a micrometer-sized spot (Figure 2b). As a result a simple disequilibrium setting comes close to implementing a Darwinian process with replication and selection in a simple chamber. Furthermore, the degradation of molecules does not lead to a dead end as molecules constantly re-accumulate. Interestingly, the trap implements a selective pressure to keep the longer of the replicated molecules in play.

In the experiment, the convection cycles the temperature in 60 seconds and the accumulation reaches a steady state within the chamber after 90 seconds. However, these numbers strongly depend on the geometries. Convection can for example reach 0.1s return times in smaller, circular chamber geometries. The latter might allow the fast cooling experiments run by Matsuno (Imai et al. 1999) in a single convection chamber. What the convection certainly allows is for the implementation of adsorption/ desorption cycles on clay particles which are too large to be efficiently accumulated. The particles periodically add and remove large polymers on their surfaces and allow for efficient catalytic activities, for example along the lines of James Ferris (Ferris and Ertem 1992, Kawamura and Ferris 1994).

It is interesting to note that even in one convection chamber, the laminar convection produces trajectories which are isolated from each other by the laminarity of the flow. Diffusion can bridge between them only at a much smaller time scale than the temperature oscillation of convection. A single chamber can thus host a number of thermal cycling conditions which interact at time scales slower than the convection itself. Therefore different species of molecules could replicate in parallel under different conditions and allow for coevolution and evolution of subpopulations.

The physical location for this temperature gradient and the pores that host convection follow the lines of Corliss, Russell, Hall, Cairns-Smith, Matsuno, Martin and others (Corliss et al. 1979, Russell et al. 1988, Imai et al. 1999, Russell and Hall 1997). They pinpointed hydrothermal systems on the deep-sea floor for the origin of life (Boyce et al. 1983, Holm 1992, Van Dover 2000, Kelley 2001). Hydrothermal venting systems create large flows of hot water (50°C to 370°C). This high temperature circulation is contrasted by 2°C cold ocean water on the outside. Between are porous mineral precipitates and rocks with a multitude of porous spaces allowing a complex network of convection chambers. This network of chambers can effectively dictate different tasks, one pore could accumulate shorter polymers while a neighboring pore give rise to large scale temperature oscillations and thereby longer polymers.

4. Disequilibrium First

In our view, most interesting are disequilibrium settings which keep the concentrations of molecules at a constantly high level or implement steady states of constant change. In modern time most life can create structures within the second law of thermodynamics by – directly or indirectly – converting high energy, low entropy solar radiation into low energy, high entropy waves. To study the origin of life, we have to search for a similar geological flux of entropy which allows the buildup of structure in a disequilibrium setting. From everyday observations it is clear that life which cannot keep itself in a state of disequilibrium(i.e. is without food and without gradients), cannot be sustained and dies.

We expect that a focus on disequilibrium conditions will foster more interdisciplinary research. In the past, the field was dominated by experiments where high concentrations of molecules were used as the entropic driving force to push the reaction into the desired (and preconceived) direction. It is a typical reflex for a chemist to grab the concentrated molecule and study its reactions in a mix and analyze the setting. Only slowly is the field realizing that it is missing many interesting and creative boundary conditions in such approaches.

Freeman Dyson (Dyson 1986) has stressed that metabolism and replication have to be combined to understand the origin of life. He saw a neglect of metabolism in a genetically focused origin of life research and made the distinction between the replication of genes and the reproduction of the cell. Making use of a metabolism, a “second beginning” occurred in which nucleotides were synthesized, genes entered in a way analogous to a parasitic disease, only later to develop the symbiotic advantage of replication and reproduction. Dyson argued that this active metabolism of the beginning already shaped the possibilities and constraints of genetic information.

We would like to point out that a disequilibrium setting must predate metabolism. A living metabolism can only be created by the structure, concentrations and specific molecules possible in an entropy flow. Otherwise it has to succumb to the cruelty of the second law of thermodynamics: equilibrium systems have to decay into a state of homogeneous high entropy. To build the metabolism that can host genes, we argue that the disequilibrium setting has to predate both. Not metabolism first or genes first is crucial to create life, but the disequilibrium is at least similarly important, if not the central ingredient to drive and originate life.

We see ideas on the origin of life converging, both on the experimental and theoretical level. For example, the Szostak group showed that the hydrothermal mechanism accumulate small molecules (Braun and Libchaber 2002, Baaske et al. 2007) and will form vesicles when applied to lipids (Budin et al. 2009). As a result, vesicles form and have the co-accumulated molecules already inside. This exactly counters our above criticism that vesicles have to be prepared at high content concentrations by using a disequilibrium setting. Similarly, joining Ferris' surface catalysis with desorption and adsorption by thermal convection is likely to generate more interesting and longer nucleotide strands. Moreover, those strands are not accessible to water without washing steps. Other converging moves have older roots. Already some time ago, Hans Kuhn was focusing on how to merge ideas from replication with those of translation (Kuhn and Waser 1982). The path we see before us is the path of integration of different approaches along the line of disequilibrium settings. We will see a number of disequilibrium, often microscale, experimental ideas from more disciplines. The experiments will shift the focus from kinetic chemical equilibrium experiments (i.e. mix and analyze) towards true steady state disequilibrium conditions. We expect many new interesting results from them.