What is Life?

The definition of life is a far-reaching issue that affects not only biology and biochemistry, but also has an impact on searching for life in the Universe. The answer to the fundamental question, ‘what is life?’ has turned out to be one of the hardest to deliver.Reference Benner ⁶ Consequently, many definitions have been proposed that can be summarized by ‘what you see depends on where you stand’. Below, we present some of these definitions.

The best-known definitions from a theoretical physics perspective come from Erwin Schrödinger and the Polish philosopher and cosmologist, Michał Heller (these and some other definitions were cited during a symposium of the Polish Academy of Science ⁷ ). The first stated that living systems self-assemble against nature’s tendency toward disorder, or entropy. Similarly, Heller said that life processes low-entropy solar energy by changing it into order and releasing disorder (heat) into space. On the other hand, biologists and biochemists pay more attention to the internal processes of cells. For example, Andrzej Legocki claims that living systems are characterized by self-processed metabolism (acquiring energy) and autoreplication ability, and Włodzimierz Sedlak stated that life is a compound set of chemical reactions and electron processes in a semiconductive environment of proteins. From a molecular biology perspective, Jan Barciszewski very briefly defines life as a minimal set of important genes. One of the best-recognized definitions from the evolutionist perspective is a National Aeronautics and Space Administration of United States working definition that states that life is a self-sustaining chemical system capable of Darwinian evolution. Evolution is underlined by Jan Kozłowski, who stated that complex biochemical structures may be considered alive if they are capable of evolution via natural selection. Finally, from a computer science perspective, life can be defined as a system carrying and processing information that is capable of replicating itself without the help of other systems that do not belong to the same type.

The multitude of definitions of life presented above clearly shows that we are still lacking a general theory of living systems. The definition of life is different depending on the field of science that utilizes it. Therefore, wide-ranging multidisciplinary research will be required before a single, commonly accepted definition can be formulated. However, regardless of the definition that we choose, we can try to find processes for the beginning of life that has evolved to the form that we observe nowadays.

Probabilistic and Deterministic Hypotheses

In one of his seminal papers, ‘A Mathematical Theory of Communication’, the American mathematician, electronic engineer and cryptographer Claude Shannon linked the concepts of information and uncertainty (measured by probability).Reference Shannon ⁸ Shannon noted that the amount of information conveyed (and the amount of uncertainty reduced) in a series of symbols is inversely proportional to the probability of occurrence of a particular event or symbol. For example, the predicted outcome of rolling a six-sided die is more improbable than the outcome of flipping a coin; therefore, it conveys more information. Shannon’s theory also implies that information increases as a sequence of characters grows but it cannot distinguish functional or message-bearing sequences from random or useless ones. The same finding applies to the biological sequences of molecules. The longer the sequence and the less probable occurrence of each of its elements, the more information can be stored in biomolecules. However, it should be kept in mind that despite carrying quantitative information, which was considered by Shannon, biomolecules should also carry qualitative information (functional ones, called specificity). Therefore, their analysis requires much more complicated steps than only considering the ordering of symbols in macromolecules.

From the perspective of this probability, we can formulate two different hypotheses of the beginning of life: the random origin (chance alone) hypothesis and the origin driven by laws of nature (deterministic) hypothesis. The hypothesis of the random origin of life assumes that some random processes have led to the beginning of life. Therefore, we can calculate the probability of the formation of life and verify whether the calculated probability is reasonable according to our knowledge of biochemistry and physics.

According to the work of Axe we can assume that the shortest functional protein is built of 150 amino acids.Reference Axe ⁹ Meyer calculated the probability of forming such a protein in completely random processes.Reference Meyer ¹⁰ He claims that if we consider that functioning proteins tolerate only left-handed (L-form) amino acids, and given that the production of the right-handed (D-form) and left-handed isomers were roughly of equal frequency in abiotic conditions, the probability of randomly attaining a protein consisting only of 150 L-form amino acids is:

Moreover, he found that the probability of creation of proper peptide bonds between any two amino acids is approximately equal to half, so the chance of connecting all of them in the chain of 150 molecules is equal to:

Thus, starting from mixtures of D-forms and L-forms, the probability of building a 150-amino-acid chain at random in which all bonds are peptide bonds and all amino acids are L-form is:

assuming that all events are independent.

This probability describes only the complexity of a molecule in Shannon’s sense. This chance is even lower if a particular variation of the 150-amino-acid chain that performs a specific function is required. Taking into consideration that there are 20 amino acids, its probability is equal to:

In fact, because some mutations in amino acid chains do not change the protein function, the cassette mutagenesis technique showed that the probability of producing any functional 150-amino-acid chain is:Reference Axe ^{9 ,} Reference Bowie, Reidhaar-Olson, Lim and Sauer ¹¹

Summing up, the probability that a 150-amino-acid compound assembled by random interactions in a prebiotic soup would be a functional protein (event X) can be estimated as:

again assuming that all events are independent.

To answer whether this is a probable event or not, we have to consider whether there are enough probabilistic resources in the Universe (Ω). We can use the universal probability bound (UPB), which is a method of objectively measuring the plausibility of any probabilistic hypothesis.Reference Dembski ^{12 –} Reference Abel ¹⁴ The UPB addresses the question of how many chances particles have had to interact since the Big Bang. The value of Ω is calculated by taking into account three factors:

1. (1) The number of seconds that have elapsed since the Big Bang, which, based on the 14 billion-year-old Universe, is assumed to be 10¹⁷ seconds.

2. (2) The number of possible state changes of particles per second which, based on the minimum amount of time required for light to traverse the Planck length, called Planck time, is assumed to be the inverse of Planck time which is 10⁴³.Reference Meyer ^{10 ,} Reference Dembski ^{12 –} Reference Abel ¹⁴

3. (3) An estimation of the number of elementary particles in the observable Universe, which is 10⁸⁰.Reference Eddington ¹⁵

Summing up, with the assumption that all events are independent, we have:

which is the number of possible observable events in the Universe. This is an equivalent to the measure of the probabilistic resources that have ever existed in the entire observable Universe. This value is much too small compared with the computed chance 1 out of 10¹⁶⁴ of obtaining the functional protein of 150 amino acids. Such calculations have persuaded many scientists to recognize the random origin hypothesis as being very improbable.

Next to the random origin hypothesis, there are deterministic hypotheses. They assume that processes that lead to the beginning of life could be driven by laws of nature: either by a biochemical predestination leading to self-organization,Reference Kenyon ¹⁶ or by external factors that could also lead to self-organization. In physics, self-organization refers to a spontaneous increase of the order of the system due to some natural processes, forces or laws, e.g. a vortex forms in a bathtub as the water swirls down the drain. Self-organization theories of the origins of life try to attribute the origins of life to physical or chemical forces or processes (laws of nature). In this case, simple monomers (amino acids, bases and sugars) arose from simpler atmospheric gases and energy. Polymers (proteins, DNA and RNA) arose from monomers and then primitive membranes formed around these polymers. Finally, primitive metabolism emerged inside these membranes as various polymers interacted chemically with one another. All these processes were driven by forces of chemical necessity, which were also responsible for the origins of life.

Prigogine and Nicolis turned their attention to the role of an external energy flowing into primitive living systems that may have played a role in the origins of biological organization.Reference Nicolis and Prigogine ¹⁷ However, their theory explains only the order of molecules (in the sense of symmetrical or repeating patterns), but cannot explain the information coding found in DNA or RNA. This approach fails to explain the issue because there is no example in nature that shows that information can be created ab initio.

Alternatively, Kauffman proposes a new model, which assumes that the emergence of metabolism preceded coding of information.Reference Kauffman ¹⁸ Kauffman’s model states that a self-reproducing metabolic system might emerge directly from a set of ‘low-specificity’ catalytic peptides and RNA molecules in a prebiotic soup, which is a set of catalytic polymers in which no single molecule reproduces itself, but the system as a whole does. However, this model also fails to explain the origins of the specified complexity of biological molecules.

Necessity alone fails to explain the origins of life, but what about chance and necessity together? This is the most probable scenario because physical and chemical forces can enhance the probability of origins of life.

RNA World Hypothesis

One of the most recognized hypotheses of the origins of life is the RNA world theory that assumes that the first forms of life were made of RNA and their synthesis was catalysed by RNA polymerase, which was also an RNA molecule.Reference Cech ¹⁹ Therefore, the RNA world hypothesis is based on the assumption that RNA molecules were capable of catalysis and replication. This concept was independently proposed by a few researchers in the 1960s based on the fact that RNA can form complex secondary structures,Reference Crick ^{20 –} Reference Gesteland, Cech and Atkins ²³ which suggests that RNA can potentially exhibit catalytic properties.Reference Woese ²⁴ Nevertheless, the term ‘RNA world’ was first used in 1986 by the Chemistry Nobel Prize winner Walter Gilbert, in a commentary on how recent observations of the catalytic properties of various forms of RNA fit with this hypothesis. The RNA world theory is a combination of the random origin hypothesis and the origin driven by laws of nature hypothesis. Chance would have played a greater role at the beginning, because suitable environmental conditions were necessary prior to initiation of the laws of nature.

Since the suggestion was first made that RNA constituted the first forms of life, more and more biological and biochemical findings seem to confirm this thesis.Reference Gesteland ^{25 ,} Reference Neveu, Kim and Benner ²⁶ The first strong evidence came from the discovery of a catalytic activity of RNA molecules, which was found in a Tetrahymena group I intron by Thomas Cech and, independently, by Sydney Altman in the RNAse P complex. For these discoveries, both researchers were awarded the Nobel Prize in Chemistry in 1989. ²⁷ The next evidence was the discovery of an RNA molecule motif, a hammerhead ribozyme, which is capable of a self-catalytic cleavage.Reference Forster and Symons ²⁸ Later on, Johnston et al. showed that RNA polymerase can catalyse its own synthesis, which also is considered as strong evidence of the RNA world hypothesis.Reference Johnston, Unrau, Lawrence, Glasner and Bartel ²⁹ In the same year, the ribosome was claimed to be a ribozyme after the discovery that the catalytic site of the ribosome is composed of RNA. The ribosome was named a living fossil of the RNA world. Moreover, many co-factors of enzymatic reactions are composed of RNA, especially these that are considered as being evolutionarily old. This finding suggests the central role of RNA in current and past biochemical pathways. Additional evidence supporting the RNA world hypothesis is the existence of viroids.Reference Diener ³⁰ Viroids are mostly plant pathogens, which consist of short stretches of highly complementary, circular, single-stranded, and non-coding RNA without a protein coat. Based on their characteristics, viroids can also be considered as RNA world remnants.Reference Flores, Gago-Zachert, Serra, Sanjuán and Elena ³¹

Recently, many researchers have conducted a wide range of laboratory experimentsReference Ruiz-Mirazo, Briones and de la Escosura ³² to support assumptions of the RNA World theory. Moreover, mathematiciansReference Eigen and Schuster ^{33 ,} Reference Schuster ³⁴ and computer scientistsReference Wang, Gu, Nguyen, Springsteen and Leszczynski ^{1 –} Reference Ma and Hu ⁵ have tried to model and analyse systems that are based upon interacting RNA chains. Laboratory experiments, as well as mathematical and computer science approaches, with special attention given to the relationship between their results, are described in the next section. However, despite extensive study, it is not clear how all the pieces of the RNA world puzzle fit together, and some pieces are still missing.

Levels of Early Pre-life Organization

We propose dividing the processes of the origins of life with special attention to the RNA world hypothesis into four levels of complexity, α, β, γ, δ, and define the transitions between them that denote the change of the level of complexity. α is the level of simple inorganic compounds that were plausibly accessible on prebiotic Earth. β is the level of nucleotides. γ is the level of polynucleotide polymers. δ is the level of interacting chains of nucleotides that are either compartmentalized or unbounded. The transitions between them are as follows: α → β denotes formation of nucleotides, β → γ denotes polymerization, and γ → δ denotes self-organization. The levels of complexity and transitions between them are presented in Figure 1.

Figure 1

The levels of complexity and the transitions between them.

As for the α → β transition, well-known Miller–Urey experiments have shown that, from simple inorganic compounds under conditions that mimic the early Earth environment, it is possible to create amino acids.Reference Miller ^{35 ,} Reference Miller and Urey ³⁶ In addition to amino acids, fatty acids have been demonstrated to be easily accessible under primordial conditions by a Fisher–Tropsch type reaction.Reference McCollom, Ritter and Simoneit ³⁷ Moreover, they were found in Carbonaceous chondrite meteorites.Reference Lawless and Yuen ³⁸ However, neither amino acids nor lipids possess properties of inheritance because they cannot serve as a template for the next generation of molecules.

Deriving ribonucleotides in a set of prebiotic conditions is a more challenging task. Juan Oro derived adenine by mixing hydrogen cyanide and ammonia in aqueous solution.Reference Oro and Kimball ³⁹ Later, the Sutherland Group demonstrated experimentally unconventional high yielding routes to pyrimidine ribonucleotides through amino-oxazoline intermediates.Reference Powner, Gerland and Sutherland ⁴⁰ The Saladino group chose another reaction scheme and concentrated on formamide as the key component of the system, which under various catalysts leads to all five nucleotides (A, G, C, T and U for RNA) and their analogues.Reference Saladino, Crestini, Pino, Costanzo and Di Mauro ⁴¹ Moreover, formamide has been recently identified as one of the most common carbon-containing molecules in the Universe.Reference Despois, Crovisier, Bockelée-Morvan and Biver ^{42 ,} Reference Adande, Woolf and Ziurys ⁴³ The reaction kinetic analysis of adenine in formamide formation has been studied with density functional theory, revealing that this mechanism is energetically favourable.Reference Wang, Gu, Nguyen, Springsteen and Leszczynski ⁴⁴ There are many possible chemical routes leading to ribonucleotides that are described in the literature. Nevertheless, the scientific community has not agreed on their plausibility in primordial conditions. What is more, intermediate states and transitions that have been recognized as the most probable have not yet been fully investigated.

With regard to the β → γ transition, the possibility of self-polymerization of nucleotide chains has been experimentally demonstrated by various researchers.Reference Zaher and Unrau ^{45 –} Reference Adamala and Szostak ⁴⁸ Researchers have also suggested that minerals have an essential role in the polymerization of polynucleotides.Reference Ferris, Hill, Liu and Orgel ⁴⁶ Potential mechanisms for linking shorter RNA chains via ligation have also been studied.Reference Costanzo, Pino, Ciciriello and Di Mauro ⁴⁷ The Nobel Prize winner in Physiology or Medicine, Jack Szostak, and his group concentrated on the polymerization of nucleotide chains and the biochemistry of fatty acids, which can constitute pre-cellular compartments.Reference Szostak ⁴⁹ They have shown that a selective advantage for protocells with the addition of phospholipids can be observed even at that simple level of organization.Reference Budin and Szostak ⁵⁰ Many researchers have also conducted experiments to develop ribozymes with various functions: peptidyl transferase ribozyme,Reference Zhang and Cech ⁵¹ ligaseReference Robertson, Hesselberth and Ellington ^{52 ,} Reference Paul and Joyce ⁵³ and nucleotide synthetase.Reference Unrau and Bartel ⁵⁴ An interesting trans-aminoacylator was described in 2013,Reference Turk, Chumachenko and Yarus ⁵⁵ having five nucleotides sufficient for reaction and being at the same time the smallest ribozyme yet discovered. Fatty acids have been shown to spontaneously form bilayer structures that are similar to a membrane consisting of phospholipids.Reference Gebicki and Hicks ^{56 –} Reference Hargreaves and Deamer ⁵⁸

The γ level itself has been studied extensively by Manfred Eigen (also a Nobel Prize winner in Chemistry) and Peter Schuster, whose biomathematics approaches led to the formulation of a quasi-species mathematical model describing the population of self-replicating polynucleotide chains.Reference Eigen and Schuster ⁵⁹ Given imprecise replication, Eigen has also put forward the idea that there is an upper threshold, named the error threshold, for the length of the polynucleotide chains which, while preserved, guarantees that the information can be sustained in the quasi-species.Reference Eigen ⁶⁰ This length has been estimated to be around 100 nucleotides, which has been considered insufficient to code the polymerase with an increased accuracy of replication. This has led to Eigen’s paradox and it has posed a problem for increasing the amount of information stored in the system.

Apart from biological experiments, there is no detailed approach to study the non-template-directed polymerization reaction of ribonucleotides. Such an approach could help to reveal the crucial aspects of the polymerization mechanism and suggest how to overcome obstacles on the way to efficient polymerization before the formation of the polymerases.

Mathematical analysis regarding the γ → δ transition has shown that there is a particular structure of self-organization, which is achievable theoretically in the quasi-species population. It can allow the system to cross the error threshold. This structure is called a hypercycle.Reference Eigen and Schuster ³³ Over the years, the hypercycle theory has experienced many reformulations and methodological approaches, including application of bioinformatics. The most notable are applications of partial differential equations,Reference Boerlijst and Hogeweg ⁶¹ cellular automataReference Takeuchi and Hogeweg ^{3 ,} Reference Boerlijst and Hogeweg ^{61 –} Reference Boerlijst and Hogeweg ⁶⁴ and the stochastic formulation of Eigen’s problem.Reference Szathmáry and Demeter ⁶⁵ In 2012, Vaidya et al. published the first experimental demonstration of the emergence of a cooperative network among fragments of ribozymes, which have the ability to self-assemble. Moreover, they showed their advantage over self-replicating cycles.Reference Vaidya ⁶⁶ Several authors have argued that the traditional hypercycle model formulation using ordinary differential equations is vulnerable to parasites.Reference Takeuchi and Hogeweg ^{3 ,} Reference Hogeweg and Takeuchi ^{62 –} Reference Szathmáry and Demeter ^{65 ,} Reference May ^{67 ,} Reference Boerlijst and Hogeweg ⁶⁸ Spatial organization has been suggested to solve this problem.Reference Takeuchi and Hogeweg ^{3 ,} Reference Boerlijst and Hogeweg ^{61 ,} Reference Boerlijst and Hogeweg ⁶⁴ At the same time, Takeuchi and Hogeweg showed, by a bioinformatics approach, that the replicase-parasite system is not only resistant to parasites but, more importantly, that parasites are responsible for the formation of traveling wave patterns and they are necessary to ensure the evolutionary stability of the system.Reference Takeuchi and Hogeweg ³ The work of Vaidya et al.Reference Vaidya ⁶⁶ suggests that selfish subsystems can be integrated with cooperative subsystems and can support their growth. However, this conclusion differs substantially from results obtained in evolutionary dynamics.Reference Nowak ⁶⁹ According to evolutionary dynamics theory, selfish molecules dominate the system, even if the growth rate of a selfish subsystem in isolation is lower than cooperative subsystems. As can be seen, despite an extensive study of the role of parasites for cooperation, a different formulation of the problem can lead to completely contrasting conclusions. Moreover, regardless of the methodology that has been used to study hypercycles, the emergence of hypercyclic structure from quasi-species has not yet been explained fully. Additionally, the explanation of the emergence of the traveling waves pattern provided by Takeuchi and Hogeweg lacks mathematical formulation.Reference Takeuchi and Hogeweg ³ The methodology applied and the employed algorithms may also influence the pattern formation. Simple rules of cellular automata may result in a pattern that differs from the results obtained from simulations performed with the use of more complex methods. The dimensionality of the system may also influence pattern formation and properties.

Another issue at this level of organization is the need for membranes. Life, as we typically define it, is cellular, and the questions are why and what kind of processes led to this. The behaviour and properties of free-floating RNA chains as well as compartmentalized RNAs have been studied by wet-lab experiments,Reference Szostak ^{49 ,} Reference Budin and Szostak ^{50 ,} Reference Mansy ⁷⁰ mathematically with group theory,Reference Markvoort, Sinai and Nowak ² and with the use of cellular automata.Reference Hogeweg and Takeuchi ⁶² Mansy et al. have shown that lipid bilayers are permeable for nucleotides.Reference Mansy ⁷⁰ But the real breakthrough came with the work of Szostak’s laboratory.Reference Szostak ⁴⁹ Zhu and SzostakReference Zhu and Szostak ⁷¹ have shown a pathway for coupled vesicle growth and division. Budin and Szostak have shown that the addition of phospholipids into a protocell’s membrane gains selective advantages for the protocell.Reference Budin and Szostak ⁵⁰ Moreover, protocells can grow as a consequence of osmotic movement of solvent molecules caused by the difference in osmotic potential between the interior of the protocells and their environment. This also suggested the existence of a direct link between vesicle growth and replication of internal molecules.Reference Schrum, Zhu and Szostak ⁷²

With his group’s theoretical approach, Nowak has proven that boundaries are advantageous for the system.Reference Markvoort, Sinai and Nowak ² However, he does not consider the spatial aspect of the problem. In this context, the crowded state in a cell requires special attention as it can alter diffusion, leading to the different behaviour of the system. Hogeweg and Takeuchi explored the crucial role of the spatiality of a replicating system by simulating both free and compartmentalized RNAs, which showed the different trajectories of these two simulations.Reference Hogeweg and Takeuchi ⁶² However, they do not conclude whether one is better than the other, and if so, why. The advantages of the trade-off between compartmentalized and unbounded systems have not yet been investigated fully.

Conclusions and Future Directions

Despite great efforts put into solving the hurdles of the origins of life, there are still unsolved mysteries at every level of the complexity of life. A leading chemist in the field of the origins of life, John Sutherland, commented on recent findings: ‘The big picture is it is not an RNA world, a peptide world, a lipid world. It only works if everything is connected.’Reference Service ⁷³ In our opinion, the case is even more complicated. Although the origins of life problem has its roots in biology, chemistry and biophysics, it also intersects with the fields of mathematics and computer science. It demands not only putting various molecules together in wet-lab experiments, but, more importantly, it demands the linkage of knowledge from diverse research approaches. Determining the exact linkage between various components is a major scientific challenge, as is the coupling of various levels of complexity. Making use of the data obtained at the lower level of complexity and relating it to the higher-level characteristics, as well as considering horizontal and vertical connections and information transfer, is still a demanding issue. Only in this way will we be able to infer the collective behaviour of the system and observe the emerging global organization that cannot be derived from the behaviour of its elements. Only in this way can we understand evolution and its beginnings. To achieve this goal, it will be necessary to take a truly interdisciplinary approach. It seems that bioinformatics may provide the perfect tools to model and test various scenarios of the origins of life.Reference Gelenbe, Seref and Xu ^{74 –} Reference Pietal, Szostak, Rother and Bujnicki ⁸⁰ Using a holistic bioinformatics approach to the problem, we will be able to unify the complex scientific landscape into a coherent model. Whether we can find answers in silico is still an open question, but it seems likely that linking efforts and knowledge from the various fields of science into the holistic bioinformatics view will be able to take us one step closer to understanding the origins of life.