1. Origin of Proteins in Life on Earth

The interface between life and non-life is difficult to define. The materials that make up biological systems may have been originally formed prebiotically on Earth or conveyed from the outer space (Oró, 1961; Chyba et al., 1990; Chyba & Sagan, 1992). Panspermia and abiogenesis are contradictory concepts. Eventually, we have to specify the origin of materials, which are related to the emergence of life, in the universe. Recent studies have proposed considerable evidence of panspermia (Hoyle & Wickramasinghe, 2000; Joseph, 2009; Joseph & Schild, 2010; Napier & Wickramasinghe, 2010). The chemicals conveyed via panspermia were possibly distributed not only on Earth but also on other planets. In that sense, life could be ubiquitous in the universe. However, life on Earth is the unique source that can be scientifically utilized for tracing the origin of life and evolution. Thus, this article mainly focuses on life on Earth.

Today, the biological systems on Earth are featured as complexes of exquisite chemical reactions. These reactions are interplays of many components of various biomolecules, and the systems themselves are produced through a long evolution over the course of ca. 4 billion years. The double helical structure of DNA, which was discovered by Watson and Crick (1953), explains the general nature of genetic information storage and expression. However, the entities that perform chemical reactions in biological systems are mostly limited to proteins.

Proteins are typically assembled from 20 amino acids (with some exceptions (Bock et al., 1991; Srinivasan et al., 2002)) through amide bonds (called "peptide bonds"). Moreover, proteins are ubiquitous and contribute structural, signalling, and catalytic activities. DNA sequences finally dictate the sequence of amino acids through mRNAs. The correspondence between a triplet of nucleotides and amino acids is known as the genetic code.

Protein biosynthesis occurs on the ribosome. The modern ribosome is a complex macromolecule that consists of more than 50 proteins and at least 3 kinds of RNAs (Moore, 1991). Furthermore, the ribosome components consist of 2 subunits (large and small), whereby peptide bond formation occurs on the large subunit, whereas the decoding of the sequences of mRNAs is performed on the small subunit (Noller, 1993). A particular location of the ribosome called the "peptidyl transferase center" (PTC) promotes the formation of new peptide bonds with two tRNAs, thereby lengthening the growing polypeptide with each cycle. Based on the structure of the ribosome, I have begun efforts to solve the mystery of the evolution of the protein biosynthetic system (Fig. 1).

2. The Structure of the Ribosome

The atomic level structure of the modern ribosome has been revealed in previous studies (Ban et al., 2000; Schluenzen et al., 2000; Wimberly et al., 2000), and the Nobel Prize in Chemistry was awarded for this work in 2009. The following studies revealed not only the structures of each of the ribosomal subunit, but also of the structure of the whole ribosome. In addition, other studies have investigated the complexes of the ribosome with mRNA, tRNA, protein factors, or antibiotics (Korostelev & Noller, 2007; Ramakrishnan, 2008; Steitz, 2008).

The structure of the large ribosomal subunit revealed important information on the mechanism and origin of the PTC. Atomic resolution structures of the Haloarcula marismortui large ribosomal subunit and its complexes with substrate analogs showed no proteins closer than approximately 18 Å to the reaction center, which was surrounded exclusively by RNA (Fig. 1b) (Ban et al., 2000; Nissen et al., 2000). In particular, the substrate analogs interacted by conserved nucleotides of the Central loop (C-loop) of domain V in 23S rRNA (Nissen et al., 2000). This feature suggests that the PTC entity is composed of only RNA molecules, and may be the actual fossil of the RNA world.

3. Mechanism of Peptide Synthesis on the Ribosome

A peptide bond is formed on the PTC with an initiator tRNA or peptidyl-tRNA in the P-site (peptidyl or donor site) and an aminoacyl-tRNA in the A-site (aminoacyl or acceptor site) (Fig. 2a). The ribosomal structure suggests many possible mechanisms for peptide bond formation. Of note, the nucleobase in 23S rRNA that is closest to PTC is A2451 (in Escherichia coli). The N1 or N3 of A2451 may abstract the proton from the α-amino group of aminoacyl-tRNA and increase the nucleophilicity of the nitrogen atom for attack on the carbonyl carbon of the ester of peptidyl-tRNA (Nissen et al., 2000) (Fig. 2b). However, previous studies have shown that large ribosomal subunits with mutated A2451 have significant peptidyl transferase activity (Polacek et al., 2001; Thompson et al., 2001; Bieling et al., 2006).

In contrast to the acid-base mechanism of A2451, the importance of the hydrogen bond network including the ribose 2′-OH group of A2451 has been indicated by the circular permutation method that allows functional group replacements to be carried out on active site 23S rRNA residues (Erlacher & Polacek, 2008). Such interactions may contribute to the proper orientation of the α-amino group for nucleophilic attack. 2 essential 2′-OH groups are related to a "proton shuttle": when the 3′-ester bond of a peptidyl-tRNA at the P site is cleaved, a proton is delivered from the adjacent 2′-OH group, which in turn receives a proton from the α-amino group of the aminoacyl-tRNA at the A site (Pech & Nierhaus, 2008) (Fig. 2c).

Based on such previous findings, the ribosome seems to be a ribozyme. Pioneering work by Noller et al. revealed that protein-depleted (up to 95%) ribosomes showed significant peptide bond forming ability, which was elucidated long before the completion of the high-resolution structure of the ribosome (Noller et al., 1992). Despite research progress, it remains unclear which portion of the PTC is truly critical for peptide bond formation. Examination of the crystal structure of the ribosome would be a good starting point in clarifying this issue and elucidating the mechanism.

Neutron-scattering analysis previously showed that the position of the L2 protein (one of the proteins composed of the large subunit) moved approximately 30 Å into the PTC when the large (50S) and small (30S) ribosomal subunits reassociated to form the whole (70S) ribosome (Willumeit et al., 2001). It is difficult to imagine the involvement of a protein residue based on the crystal structure, but the L2 protein may play a direct catalytic role in peptide bond formation (i.e., the histidine-catalyzed acid-base mechanism) (Cooperman et al., 1995) because biomolecular interactions pertain to the dynamic nature of the structure.

4. Origin of Peptide Bond Formation

In considering the peptide bond formation of the ribosome, it is necessary to contemplate how the complex machinery has come about. Prebiotic amino acid formation is highly plausible, as previously demonstrated by an experiment performed by Miller (Miller, 1953). Moreover, it was reported recently that activated pyrimidine ribonucleotides are formed in a short sequence through arabinose amino-oxazoline and anhydronucleoside intermediates (Powner et al., 2009). The establishment of biosynthetic pathways in primitive biological systems may have added to the number of amino acids used in the systems, and the genetic code may have been established gradually (Wong, 1975).

The second law of thermodynamics indicates that peptide bond formation does not occur spontaneously. Therefore, energy must be added into the system by some means and amino acids must be "activated." Modern biological systems use the energy of the ATP hydrolysis for coupling many reactions (Lipmann, 1941). However, during the prebiotic stage, the light from the sun, geothermal energy, pressure in the thermal vent, or other similar sources may have been used in the process of activating the molecules of a system. The development of prebiotic precursors of biomolecules might have occurred in interstellar space, and were subsequently transferred to Earth by comets, asteroids, or meteorites (Oró, 1961; Chyba et al., 1990; Chyba & Sagan, 1992). Reactions on clay (Paecht-Horowitz et al., 1970) and/or dry mixtures of amino acids (Fox & Harada, 1958) may have facilitated the condensation of activated amino acids, thereby forming peptide bonds. Iron sulfate is known to cause unusual reducing reactions, especially with H₂S. Wächtershäuser (1992) previously proposed the idea of an "iron-sulfur world" where low-molecular weight constituents may have originated autotrophic metabolism. In such circumstances, amino acids would have been converted into simple peptides (Huber & Wächtershäuser, 1998). In fact, it has been demonstrated that the peptide containing a thioester at the carboxyl terminal undergoes nucleophilic attack by the side chain of the Cys residue at the amino terminal of another peptide. Moreover, the formed thioester ligation product readily undergoes a rapid intramolecular reaction at the α-amino group of the Cys to yield a product with a native peptide bond. This series of events is called "native chemical ligation" and is important in the general application of protein chemistry (Dawson et al., 1994). These possibilities should be further considered in terms of the very early mechanisms responsible for peptide bond formation. However, because we must consider the modern ribosome, we cannot avoid consideration of RNA in the evolution of biological systems.

5. Emergence of the RNA World

The peptide synthesis hypotheses described in the section above must jump over a large gap to attain ribosome-based peptide synthesis. Christian de Duve proposed the "thioester world" hypothesis (de Duve, 1995) to fill this gap. Thioesters are high energy intermediates that are still used by modern biological systems (e.g., acetyl-CoA). The energy derived from thioester hydrolysis may be used to form peptide bonds. In fact, non-ribosomal peptide synthesis that occurs in current systems is performed using this rationale. The enzymes involved in non-ribosomal peptide synthesis are apparently highly evolved and are highly systematic compounds with strictly sequential functional domains (Marahiel et al., 1997). The thioester world might be a potential link between the proposed iron-sulfur world and RNA world. In modern biological systems, DNA is both used to generate proteins and generated by proteins. This chicken-or-egg conundrum, or in other words whether the nucleic acids or proteins originated first, seems to have been solved by the previous discoveries of both Cech’s and Altman’s groups, which showed that RNA not only carries genetic information but also functions as a catalyst (Kruger et al., 1982; Guerrier-Takada et al., 1983). This led Gilbert to propose the "RNA world" hypothesis to describe the possible initial stage of life (Gilbert, 1986).

Peptide bond formation based on the proximity effect using nucleotides and oligonucleotides can be useful. If the activated amino acid is attached on a single RNA, another activated amino acid on another RNA can easily come together with the first RNA using complementary base pairing, which finally forms a peptide bond. Based on these strategies, several successful attempts have been carried out in previous studies (Weber & Orgel, 1978; 1979).

The RNA world demands that peptide bond formation be catalyzed by ribosomal RNA alone, however, this has not been successfully demonstrated in the laboratory (Khaitovich et al., 1999). On the other hand, in vitro selection methods have isolated a ribozyme capable of catalyzing peptide bond formation with 196 nucleotides (Zhang & Cech, 1997), although spontaneous formation of the ribozyme with such a length is quite unbelievable.

6. Minihelix and tRNA

tRNA is typically composed of 76 nucleotides. If all combinations of the nucleotide sequences were tested during the primitive elongation process of forming full-length tRNA (4⁷⁶), the mass of the required materials would be about one-hundredth of the whole Earth. This calculation implies that full-length tRNA may have been produced at least in part by the duplication of the small parts of tRNA. A minihelix is just a half-sized molecule of tRNA and corresponds to one arm of the L-shaped three-dimensional structure of tRNA (Fig. 1a). Minihelices can act as the substrates of many aminoacyl tRNA synthetases, which are key enzymes that catalyze amino acid attachment to tRNAs. The other arm of the L-shaped structure of tRNA decodes the incoming mRNA using codon-anticodon interactions. The 2 arms can be physically dissected to investigate the possible evolutionary origins of these individual functions. A minihelix also retains the native function of peptide synthesis on the ribosome, and is the ancient part of tRNA (Fig. 1a) (Schimmel et al., 1993; Schimmel & Ribas de Pouplana, 1995).

7. Reaction with Aminoacyl Phosphate Adaptors

In the aminoacylation of tRNA in the current system, it is noteworthy that aminoacyl tRNA synthetases synthesize aminoacyl adenylate in the first step (Schimmel, 1987). As aminoacyl adenylate is also formed under prebiotic conditions (Paecht-Horowitz & Katchalsky, 1973), aminoacyl phosphate oligonucleotide, which has the same linkage as that of aminoacyl adenylate, would have been formed in the RNA world. With this in mind, we attempted template-directed peptide synthesis in a system consisting of an aminoacyl-minihelix, aminoacyl phosphate oligonucleotide and a template-like guide RNA, which revealed the elongation continued over step 2, as shown in Fig. 3a (Tamura & Schimmel, 2003; Tamura & Alexander, 2004). In addition, aminoacyl phosphate oligonucleotide performed chiral selective aminoacylation of the minihelix with a template-like guide RNA (Tamura & Schimmel, 2004), which is likely an important clue to the origin of the homochirality of biological systems (Tamura & Schimmel, 2006; Tamura, 2008; 2009; 2010).

8. Focus on the Interaction of CCA and 23S rRNA

Ribosome-catalyzed peptidyl transferase activity of the PTC depends on base pairing between 23S rRNA nucleotides and the universal CCA sequence of tRNA (Moazed & Noller, 1991; Tamura, 1994; Nissen et al., 2000). The single-stranded CCA sequence is conserved among all tRNAs (Fig. 1a), and amino acids are attached at the OH group of adenosine through ester bond linkage during aminoacylation. The origin of this sequence is thought to be a genomic tag of a self-replicating ribozyme in a putative RNA world (Weiner & Maizels, 1987). A simplified system using an artificial RNA with 5′-5′ phosphodiester linkage to represent the minimum essence of both aminoacyl-tRNA and rRNA has also been investigated (Fig. 3b). An aminoacylated minihelix (N-acetylalanyl-minihelix) mixed with a puromycin containing oligo RNA (Pm-UGGU), where puromycin is an analog of aminoacyl-tRNA, produced N-acetylalanyl-puromycin product (i.e., peptide bond formation). Of note, peptide bond formation was dependent on the base complementarity between the Pm-UGGU and the minihelix 3′-CCA (Tamura & Schimmel, 2001).

9. Possible Evolutionary Route to the Ribosome

The aforementioned information reveals the possible involvement of primitive tRNA (like RNA minihelix) in peptide bond formation in the primordial system. However, the ultimate route to the ribosome remains unclear. With this information in mind, the ribosome structure can be reevaluated. The structure of the large subunit of Deinococcus radiodurans may provide a new concept about the origin of the PTC. PTC is formed by a pocket-like symmetrical RNA dimer that is composed of two L-shaped RNA units (Agmon et al., 2005; Agmon, 2009), which were likely proto-ribosomes (Fig. 4). The symmetrical association of the 2 units may have provided the proper positioning of the reactants (Agmon et al., 2005; Agmon, 2009), leading to a stereochemistry favorable for peptide bond formation to occur in a similar way to that in the modern ribosome. The size of the each symmetrical dimer is approximately the same as that of tRNA. In the RNA world, the maximum possible elongated size of the first reproductive units would be up to that of tRNA (~76mer), upon consideration of the hypercycle theory by Eigen (Eigen & Schuster, 1977), which is also related to the error-catastrophe theory by Orgel (Orgel, 1963). It is plausible that tRNA would have been a primordial molecule composed of proto-ribosomes. The symmetrical RNA dimer composed of two L-shaped RNA units would have been a scaffold that enabled precise positioning of 2 aminoacyl-tRNAs (Agmon et al., 2005; Agmon, 2009).

Experimental verification of this idea should be attempted in future studies. Proto-ribosomes may have started functioning in the RNA world, but formed proteins would have likely been involved in the evolution of the modern ribosome. The concept of structural mimicry between translational protein factors and tRNA could fill the missing link of the evolution of the ribosome (Ito et al., 1996). Release factors mimic the tRNA structure and recognize stop codons of mRNA. The phenomenon related to "peptide anticodon" (Ito et al., 2000) could have been the crucial relic of the transition from the RNA world to the protein world.

10. Understanding Life from the Cosmological Standpoint

Proteins are indispensable in the biological life system on Earth. Proteins are composed of amino acids, which would be ubiquitous in the universe: in nebular cloud, in a comet, and/or on a Super-Earth in another solar system. Thus, the so-called "seeds" of life are present everywhere in the universe. Panspermia, especially at a material level, seems possible, and considerable evidence has been provided to validate this concept (Crick & Orgel, 1973; Hoyle & Wickramasinghe, 2000; Joseph, 2009; Joseph & Schild, 2010; Napier & Wickramasinghe, 2010). In our solar system, there is a good chance of existence of water in Europa (Kivelson et al., 2000); therefore, the existence of life on such satellites (or planets) seems theoretically plausible. The establishment of protein-based life on the present Earth must have overcome two obstacles.

First, the units composed of biomolecules might have polymerized to function as the "building blocks" of life and facilitate the interactions among the blocks. These units contain peptides and nucleotides of specific lengths. These productions and polymerizations might have occurred under special conditions on Earth, such as gravity, geothermal energy, hydrothermal energy, and the tilt of Earth’s axis; these factors possibly make Earth a unique bed of life. In contrast, panspermia might have just contributed to the transfer of relatively small molecules.

The second obstacle must have been self-organization of the molecules. Biomolecular organization has many levels, and in the evolutionary sense, the foundation of peptide bond formation machinery, which is the point of discussion in this article, is most critical from the viewpoint of the origin of life. Undoubtedly, the prototype of the machinery (proto-ribosome) was formed during the repetitive process of relatively small building block formation. The role of RNA in the evolutionary process related to the development of ribosome is unquestionable, and the appearance and ability of the ribosome may also be specific to Earth, at least to our knowledge.

The question that should definitely be discussed in this article is "Is life on Earth, especially the system of the protein-forming machinery, the ribosome, really unique to Earth?" The answer probably would be "no" because recent astronomical studies have suggested that terrestrial planets are not rare entities in the universe (Ida & Lin, 2004). Based on the standard model of planetary system formation, more than 10% of all solar-type stars in our galaxy could be orbited by habitable planets (Ida, 2005).

For the emergence of life, several hurdles along the process of evolution must have been overcome. Panspermia could have played an important role in the transfer of chemicals to Earth, and must have finally constituted bio-building blocks in the course of time. RNA has definitely contributed the evolution of life on Earth, and it eventually produced the ribosome. Thus, life appeared only after confronting such singularities.

11. Conclusion

Proteins are key molecules that catalyze exquisite reactions in biological systems. Therefore, the elucidation of the origin of the peptide bond-forming machinery, the ribosome, is quite important in understanding "life" itself. High-resolution crystal structures have provided new insights into the mechanism of peptide bond formation in biological systems. However, there remains a large gap concerning the events that occurred between ribosomal peptide synthesis and the model of simple molecules presented to date. Future research will no doubt shed light on this gap. I propose that the key to this research would be to determine how small RNAs can act as proto-ribosomes.

Acknowledgments: I thank Dr. Koichi Ito for valuable comments on the manuscript and for support in assembly of figures. I also thank Dr. Christopher A. Myers for helpful discussion. This work was supported by a grant from PRESTO (RNA and Biofunctions), the Japan Science and Technology Agency (JST), Japan, and by the program for development of strategic research center in private universities by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.