The origin of the genetic code is one of the most fundamental and difficult problems in biology (Schimmel 1996). Many attempts have been made to identify the stereochemical interactions between amino acids and their coding nucleotides, but without complete/much success (Fontecilla-Camps 2014; Komatsu et al. 2014; Lacey and Mullins 1983; Shimizu 1982; Yarus 1998). In a recent issue of J. Mol. Evol., Bernhardt and Patrick (2014) proposed the possibility that the genetic code originated with the incorporation of glycine (Gly). Focusing on the second and third nucleotides of the anticodon of tRNAGly (C35C36) and the following adenosine (A37), it has been hypothesized that the minihelix-like hairpin RNA possessing a CCA terminus was ligated in tandem to evolve to full length-tRNAGly (Bernhardt and Tate 2008, 2010) (Fig. 1a). This author here proposes a new model of the primordial glycylation of tRNAGly.

Fig. 1
figure 1

Origin of tRNAGly and glycylation of the minihelix. a Full-length tRNAGly is produced by tandem-ligation of a minihelix-like hairpin monomer with the NCCA-3′ terminal. The canonical intron insertion point corresponds exactly to the joint position (between positions 37 and 38, as shown in the triangle). b The minihelix with a UCCA-3′ terminal forms a folded-back structure, and the adenine of glycyl-AMP can base-pair with the uridine in the UCCA-3′ of the minihelix, accomplishing the glycylation (left). In contrast, the minihelix with an ACCA-3′ terminal forms an upright structure in which the 3′-sequence extends to the stacking of the acceptor stem (right)

Full size image

tRNA aminoacylation is performed by aminoacyl-tRNA synthetases via aminoacyl-AMP formation (Schimmel 1987). There is evidence demonstrating that aminoacyl-AMPs were formed prebiotically (Paecht-Horowitz and Katchalsky 1973). The chemistry of this reaction should be amino acid independent, in which there is a nucleophilic attack of the α-carboxylate anion of amino acids on the α-phosphorus of ATP. tRNA aminoacylation from aminoacyl-AMP is a thermodynamically downhill reaction (approximately 3 kcal/mol) (Carpenter 1960). Therefore, in principle, the oxygen of the terminal 3′-OH (or 2′-OH) of tRNA would attack the carbonyl carbon in the acyl phosphate of aminoacyl-AMP, producing an aminoacyl ester bond between tRNA and the amino acid. Although the aminoacyl phosphate linkage is extremely labile and is hydrolyzed very quickly, this rationale has been verified experimentally (Tamura and Schimmel 2004, 2006).

A minihelix with a UCCA-3′ terminal forms a structure in which the 3′-sequence folds back such that the 3′-terminal A is in close proximity to the 5′-terminal G (Puglisi et al. 1994) (Fig. 1b), whereas a minihelix with an ACCA-3′ terminal forms a structure in which the 3′-sequence extends the A-form stacking of the acceptor stem (Limmer et al. 1993; Puglisi et al. 1994) (Fig. 1b). Therefore, it is possible that the UCCA-3′ allows for internal A–U base pairing (Puglisi et al. 1994). However, the riboses of UCCA-3′ adopt mostly a 2′-endo conformation, and the stacking of these bases is not strong enough. The free energy for one hydrogen bond is approximately 0.7–1.6 kcal/mol (Kool 2001), and thermal motion can easily disrupt the plausible A–U base pairing. Therefore, given the plausible/possible existence of glycyl-AMP, the A of glycyl-AMP should also base-pair with the U of UCCA-3′ (Fig. 1b). Stacking between the A of the glycyl-AMP and the 5′-terminal G at the end of the double helix may assist with the base paring. The importance of the stacking of mononucleotides is also demonstrated by the fact that activated AMP can be oligomerized using a poly(U) template (Osawa et al. 2005).

Eubacterial tRNAGlys have a terminal UCCA-3′ (with U73 at the discriminator position), although both archaebacterial and eukaryotic cytoplasmic counterparts have an ACCA-3′ terminal (with A73) (Jühling et al. 2009; Chien et al. 2014). Remarkably, the U73-containing tRNAGlys are commonly recognized by eubacterial glycyl-tRNA synthetases with different quaternary structures (e.g., α2β2-type in E. coli and α2-type in T. thermophilus) (Mazauric et al. 1996; Nameki et al. 1997). Therefore, there may be two distinct evolutionary origins depending on the discriminator (Chien et al. 2014).

Thus, prebiotically formed glycyl-AMP could have base-paired with the uridine in the folded-back UCCA-3′ of the minihelix, and then the Gly residue would transfer to the OH group of the 3′-terminal adenosine, thereby accomplishing glycylation (Fig. 1b). A DNA-templated model reaction system also demonstrated that chemical reactions efficiently occur at the location with a 5-base interval from the reaction site (Calderone and Liu 2004), suggesting that the flexibility of the folded-back UCCA-3′ is not unfavorable for the reaction.

tRNACys from all three kingdoms and tRNAGln from the eukaryotic cytoplasm also possesses a U73 (Jühling et al. 2009), but the role of these two tRNAs could be ruled out for the present discussion. This is because in the biosynthetic pathways, Gln and Cys are produced from Glu and Ser, respectively, and both Gln and Cys are suggested to have been incorporated in the genetic code later in evolution, according to the coevolution theory (Wong 1975). In addition, prokaryotes plausibly appeared before eukaryotes. Furthermore, the enzymes of aerobic thermophiles tend to contain fewer Cys residues, and there is no Cys in thermophilic isopropylmalate dehydrogenase (Kagawa et al. 1984; Ohnuma et al. 2011). In fact, the Cβ–Sγ bond of Cys residues is labile at high temperature under aerobic conditions (Jaenicke and Bohm 1998). In contrast, Gly is the simplest and most abundant amino acid, and could have been formed in primitive Earth or even in space (Kvenvolden et al. 1970; Miller 1953).

The present model is not stereochemical in the strict sense, but also does not contradict Eigen’s theory that Gly, Ala, Asp, and Val were the first amino acids designated by the GNC code (Eigen and Schuster 1977). The codon assignments of the genetic code seem to have evolved in a manner that minimizes the effects of mutations, i.e., similar amino acids are assigned in close positions (Crick 1968). Early peptides produced containing Gly would probably have been short and water-soluble, although some other amino acids that are mutually interconvertible in biosynthetic pathways (e.g., Ala, Asp, Glu, Gly, and Ser) could have also been incorporated into early proteins. Orgel (1977) pointed out the importance of β-turns stabilized by β-sheets as plausible active sites of early enzymes. Gly and Pro are often used as components of a β-turn, due to Gly’s flexibility and Pro’s cyclic structure ideally suited for the formation of the turn (Jurka and Smith 1987; Komatsu et al. 2014; Wilmot and Thornton 1988). Gly affords flexibility to peptides and might have been involved in the early coding system. The proposed mechanism for the assignment of Gly to tRNAGly (containing UCCA-3′) offers an entirely new perspective on the origin of the genetic code beyond “the frozen accident”.