Synthesis and Aminoacylation of 3′-Amino-3′-deoxy Transfer RNA and Its Activity in Ribosomal Protein Synthesis

3'-amino-3'-deoxy adenosine was enzymatically converted into 3'-amino-3'-deoxy ATP. This ATP analogue was incorporated into the 3'-terminal adenosine position of Escherichia coli tRNA. The modified tRNA was aminoacylated with phenylalanine by use of E. coli phenylalanyl tRNA synthetase. The phenylalanine was attached to the 3'-amino group of the tRNA, as shown by its high resistance to base-catalyzed hydrolysis in contrast with the normal lability of phenylalanyl-tRNA. Aminoacyl tRNA synthetases charge the 3'-amino-3'-deoxy tRNA with kinetics that are similar to those of the charging reaction in which normal tRNA is the substrate. When phenylalanyl-3'-amino-3'-deoxy tRNA is used in a protein-synthesizing system directed by poly(U) in vitro, this molecule is capable of receiving an acetyl-phenylalanine from the donor site of the ribosome. However, the ribosome is unable to cleave the amide bond connecting phenylalanine to the tRNA molecule; hence, the phenylalanyl-3'-amino-3'-deoxy tRNA has acceptor but not donor activity in protein synthesis. Failure of the ribosome to cleave the amide bond may be due to its greater stability relative to the normal ester bond. However, it may also be due to the fact that the isomerization of the peptidyl chain between the 2' and 3' positions of adenosine is prevented due to the amide bond at the 3' position, and cleavage may normally occur with the peptidyl chain on the 2' position of adenosine.     

Site of aminoacylation of tRNAs from Escherichia coli with respect to the 2'- or 3'-hydroxyl group of the terminal adenosine.

A method is presented by which the site of primary attachment of the amino acids with respect to the 2'- or 3'-hydroxyl group of the terminal adenosine of E. coli tRNAs can be determined. It is found that the aminoacyl-tRNA synthetases (EC 6.1.1.-) with specificity for Arg, Asn, Ile, Leu, Met, Phe, Thr, Trp, and Val attach the amino acid to the 2'-position; those with specificity for Gly, His, Lys, and Ser attach the amino acid to the 3'-position; and that Tyr and Cys can be enzymatically attached to both the 2'- and 3'-positions. Together with previous experiments on yeast aminoacyl-tRNA synthetases, it is now shown that the specificity for one particular hydroxyl group is preserved during the evolution from prokaryotic to eukaryotic systems.     

Progress toward the evolution of an organism with an expanded genetic code

Several significant steps have been completed toward a general method for the site-specific incorporation of unnatural amino acids into proteins in vivo. An "orthogonal" suppressor tRNA was derived from Saccharomyces cerevisiae tRNA2Gln. This yeast orthogonal tRNA is not a substrate in vitro or in vivo for any Escherichia coli aminoacyl-tRNA synthetase, including E. coli glutaminyl-tRNA synthetase (GlnRS), yet functions with the E. coli translational machinery. Importantly, S. cerevisiae GlnRS aminoacylates the yeast orthogonal tRNA in vitro and in E. coli, but does not charge E. coli tRNAGln. This yeast-derived suppressor tRNA together with yeast GlnRS thus represents a completely orthogonal tRNA/synthetase pair in E. coli suitable for the delivery of unnatural amino acids into proteins in vivo. A general method was developed to select for mutant aminoacyl-tRNA synthetases capable of charging any ribosomally accepted molecule onto an orthogonal suppressor tRNA. Finally, a rapid nonradioactive screen for unnatural amino acid uptake was developed and applied to a collection of 138 amino acids. The majority of glutamine and glutamic acid analogs under examination were found to be uptaken by E. coli. Implications of these results are discussed.     

Aminoacyl-tRNA synthetases catalyze AMP----ADP----ATP exchange reactions, indicating labile covalent enzyme-amino-acid intermediates

Aminoacyl-tRNA synthetases (amino acid-tRNA ligases, EC 6.1.1.-) catalyze the aminoacylation of specific amino acids onto their cognate tRNAs with extraordinary accuracy. Recent reports, however, indicate that this class of enzymes may play other roles in cellular metabolism. Several aminoacyl-tRNA synthetases are herein shown to catalyze the AMP----ADP and ADP----ATP exchange reactions (in the absence of tRNAs) by utilizing a transfer of the gamma-phosphate of ATP to reactive AMP and ADP intermediates that are probably the mixed anhydrides of the nucleotide and the corresponding amino acid. AMP and ADP produce active intermediates with amino acids by entering the back-reaction of amino acid activation, reacting with labile covalent amino acid-enzyme intermediates. Gramicidin synthetases 1 and 2, which are known to activate certain amino acids through the formation of intermediate thiol-esters of the amino acids and the enzymes, catalyze the same set of reactions with similar characteristics. Several lines of evidence suggest that these activities are an inherent part of the enzymatic reactions catalyzed by the aminoacyl-tRNA synthetases and gramicidin synthetases and are not due to impurities of adenylate kinase, NDP kinase, or low levels of tRNAs bound to the enzymes. The covalent amino acid-enzyme adducts are likely intermediates in the aminoacylation of their cognate tRNAs. The use of gramicidin synthetases has thus helped to illuminate mechanistic details of amino acid activation catalyzed by the aminoacyl-tRNA synthetases.     

Structural plasticity of an aminoacyl-tRNA synthetase active site

Recently, tRNA aminoacyl-tRNA synthetase pairs have been evolved that allow one to genetically encode a large array of unnatural amino acids in both prokaryotic and eukaryotic organisms. We have determined the crystal structures of two substrate-bound Methanococcus jannaschii tyrosyl aminoacyl-tRNA synthetases that charge the unnatural amino acids p-bromophenylalanine and 3-(2-naphthyl)alanine (NpAla). A comparison of these structures with the substrate-bound WT synthetase, as well as a mutant synthetase that charges p-acetylphenylalanine, shows that altered specificity is due to both side-chain and backbone rearrangements within the active site that modify hydrogen bonds and packing interactions with substrate, as well as disrupt the alpha8-helix, which spans the WT active site. The high degree of structural plasticity that is observed in these aminoacyl-tRNA synthetases is rarely found in other mutant enzymes with altered specificities and provides an explanation for the surprising adaptability of the genetic code to novel amino acids.     

Position of aminoacylation of individual Escherichia coli and yeast tRNAs.

Transfer RNAs terminating 2'-or 3'-deoxyadenosine were prepared from unfractionated E. coli and yeast (Saccharomyces cerevisiae) tRNAs and purified to remove unmodified tRNAs. The modified tRNA species were assayed for aminoacylation with each of the 20 amino acids to determine the initial position of tRNA aminoacylation. The E. coli and yeast aminoacyl-tRNA synthetases specific for arginine, isoleucine, leucine, methionine, phenylalanine, and valine, as well as the E. coli glutamyl-tRNA synthetase, aminoacylated only those cognate tRNAs terminating in 3'-deoxyadenosine (i.e., those having a 2'-OH group). On the other hand, those E. coli and yeast synthetases specific for alanine, glycine, histidine, lysine, proline, serine, and threonine, as well as the yeast synthetase specific for glutamine, utilized exclusively those tRNAs having an available 3'-OH group on the 3'-terminal nucleoside, while the E. coli and yeast synthetases specific for asparagine, cysteine, and tyrosine, and the yeast aspartyl-tRNA synthetase, utilized both of the modified cognate tRNAs. The only observed difference in specificity between the E. coli and yeast systems was for tRNATrp, which was aminoacylated on the 2'-position in E. coli and the 3'-position in yeast. The results indicate that the initial position of aminoacylation is not uniform for all tRNAs, although for individual tRNAs the specificity has been conserved during the evolution from a prokaryotic to eukaryotic organism.     

Hydrolytic editing by a class II aminoacyl-tRNA synthetase

Editing reactions catalyzed by aminoacyl-tRNA synthetases are critical for accurate translation of the genetic code. To date, this activity, whereby misactivated amino acids are hydrolyzed either before or after transfer to noncognate tRNAs, has been characterized extensively only in the case of class I synthetases. Class II synthetases have an active-site architecture that is completely distinct from that of class I. Thus, findings on editing by class I synthetases may not be applicable generally to class II enzymes. Class II Escherichia coli proline-tRNA synthetase is shown here to misactivate alanine and to hydrolyze the noncognate amino acid before transfer to tRNA(Pro). This enzyme also is capable of rapidly deacylating a mischarged Ala-tRNA(Pro) variant. A single cysteine residue (C443) that is located within the class II-specific motif 3 consensus sequence was shown previously to be dispensable for proline-tRNA synthetase aminoacylation activity. We show here that C443 is critical for the hydrolytic editing of Ala-tRNA(Pro) by this class II synthetase.     

Rational design of an evolutionary precursor of glutaminyl-tRNA synthetase

The specificity of most aminoacyl-tRNA synthetases for an amino acid and cognate tRNA pair evolved before the divergence of the three domains of life. Glutaminyl-tRNA synthetase (GlnRS) evolved later and is derived from the archaeal-type nondiscriminating glutamyl-tRNA synthetase (GluRS), an enzyme with relaxed tRNA specificity capable of forming both Glu-tRNA(Glu) and Glu-tRNA(Gln). The archaea lack GlnRS and use a specialized amidotransferase to convert Glu-tRNA(Gln) to Gln-tRNA(Gln) needed for protein synthesis. We show that the Methanothermobacter thermautotrophicus GluRS is active toward tRNA(Glu) and the two tRNA(Gln) isoacceptors the organism encodes, but with a significant catalytic preference for tRNA(Gln2)(CUG). The less active tRNA(Gln1)(UUG) responds to the less common CAA codon for Gln. From a biochemical characterization of M. thermautotrophicus GluRS variants, we found that the evolution of tRNA specificity in GlnRS could be recapitulated by converting the M. thermautotrophicus GluRS to a tRNA(Gln) specific enzyme, solely through the addition of an acceptor stem loop present in bacterial GlnRS. One designed GluRS variant is also highly specific for the tRNA(Gln2)(CUG) isoacceptor, which responds to the CAG codon, and shows no activity toward tRNA(Gln1)(UUG). Because it is now possible to eliminate particular codons from the genome of Escherichia coli, additional codons will become available for genetic code engineering. Isoacceptor-specific aminoacyl-tRNA synthetases will enable the reassignment of more open codons while preserving accurate encoding of the 20 canonical amino acids.     

Engineering a tRNA and aminoacyl-tRNA synthetase for the site-specific incorporation of unnatural amino acids into proteins in vivo

In an effort to expand the scope of protein mutagenesis, we have completed the first steps toward a general method to allow the site-specific incorporation of unnatural amino acids into proteins in vivo. Our approach involves the generation of an “orthogonal” suppressor tRNA that is uniquely acylated in Escherichia coli by an engineered aminoacyl-tRNA synthetase with the desired unnatural amino acid. To this end, eight mutations were introduced into tRNA₂^Gln based on an analysis of the x-ray crystal structure of the glutaminyl-tRNA aminoacyl synthetase (GlnRS)–tRNA₂^Gln complex and on previous biochemical data. The resulting tRNA satisfies the minimal requirements for the delivery of an unnatural amino acid: it is not acylated by any endogenous E. coli aminoacyl-tRNA synthetase including GlnRS, and it functions efficiently in protein translation. Repeated rounds of DNA shuffling and oligonucleotide-directed mutagenesis followed by genetic selection resulted in mutant GlnRS enzymes that efficiently acylate the engineered tRNA with glutamine in vitro. The mutant GlnRS and engineered tRNA also constitute a functional synthetase–tRNA pair in vivo. The nature of the GlnRS mutations, which occur both at the protein–tRNA interface and at sites further away, is discussed.     

Interactions between tRNA identity nucleotides and their recognition sites in glutaminyl-tRNA synthetase determine the cognate amino acid affinity of the enzyme.

Sequence-specific interactions between aminoacyl-tRNA synthetases and their cognate tRNAs both ensure accurate RNA recognition and prevent the binding of noncognate substrates. Here we show for Escherichia coli glutaminyl-tRNA synthetase (GlnRS; EC 6.1.1.18) that the accuracy of tRNA recognition also determines the efficiency of cognate amino acid recognition. Steady-state kinetics revealed that interactions between tRNA identity nucleotides and their recognition sites in the enzyme modulate the amino acid affinity of GlnRS. Perturbation of any of the protein-RNA interactions through mutation of either component led to considerable changes in glutamine affinity with the most marked effects seen at the discriminator base, the 10:25 base pair, and the anticodon. Reexamination of the identity set of tRNA(Gln) in the light of these results indicates that its constituents can be differentiated based upon biochemical function and their contribution to the apparent Gibbs' free energy of tRNA binding. Interactions with the acceptor stem act as strong determinants of tRNA specificity, with the discriminator base positioning the 3' end. The 10:25 base pair and U35 are apparently the major binding sites to GlnRS, with G36 contributing both to binding and recognition. Furthermore, we show that E. coli tryptophanyl-tRNA synthetase also displays tRNA-dependent changes in tryptophan affinity when charging a noncognate tRNA. The ability of tRNA to optimize amino acid recognition reveals a novel mechanism for maintaining translational fidelity and also provides a strong basis for the coevolution of tRNAs and their cognate synthetases.     

A truncated aminoacyl-tRNA synthetase modifies RNA

Aminoacyl-tRNA synthetases are modular enzymes composed of a central active site domain to which additional functional domains were appended in the course of evolution. Analysis of bacterial genome sequences revealed the presence of many shorter aminoacyl-tRNA synthetase paralogs. Here we report the characterization of a well conserved glutamyl-tRNA synthetase (GluRS) paralog (YadB in Escherichia coli) that is present in the genomes of >40 species of proteobacteria, cyanobacteria, and actinobacteria. The E. coli yadB gene encodes a truncated GluRS that lacks the C-terminal third of the protein and, consequently, the anticodon binding domain. Generation of a yadB disruption showed the gene to be dispensable for E. coli growth in rich and minimal media. Unlike GluRS, the YadB protein was able to activate glutamate in presence of ATP in a tRNA-independent fashion and to transfer glutamate onto tRNA(Asp). Neither tRNA(Glu) nor tRNA(Gln) were substrates. In contrast to canonical aminoacyl-tRNA, glutamate was not esterified to the 3'-terminal adenosine of tRNA(Asp). Instead, it was attached to the 2-amino-5-(4,5-dihydroxy-2-cyclopenten-1-yl) moiety of queuosine, the modified nucleoside occupying the first anticodon position of tRNA(Asp). Glutamyl-queuosine, like canonical Glu-tRNA, was hydrolyzed by mild alkaline treatment. Analysis of tRNA isolated under acidic conditions showed that this novel modification is present in normal E. coli tRNA; presumably it previously escaped detection as the standard conditions of tRNA isolation include an alkaline deacylation step that also causes hydrolysis of glutamyl-queuosine. Thus, this aminoacyl-tRNA synthetase fragment contributes to standard nucleotide modification of tRNA.     

Effector region of the translation elongation factor EF-Tu.GTP complex stabilizes an orthoester acid intermediate structure of aminoacyl-tRNA in a ternary complex.

tRNA(Val) from Escherichia coli was aminoacylated with [1-13C]valine and its complex with Thermus thermophilus elongation factor EF-Tu.GTP was analyzed by 13C NMR spectroscopy. The results suggest that the aminoacyl residue of the valyl-tRNA in ternary complex with bacterial EF-Tu and GTP is not attached to tRNA by a regular ester bond to either a 2'- or 3'-hydroxyl group; instead, an intermediate orthoester acid structure with covalent linkage to both vicinal hydroxyls of the terminal adenosine-76 is formed. Mutation of arginine-59 located in the effector region of EF-Tu, a conserved residue in protein elongation factors and the alpha subunits of heterotrimeric guanine nucleotide-binding regulatory proteins (G proteins), abolishes the stabilization of the orthoester acid structure of aminoacyl-tRNA.     

Mechanism of tRNA-dependent editing in translational quality control

Protein synthesis requires the pairing of amino acids with tRNAs catalyzed by the aminoacyl-tRNA synthetases. The synthetases are highly specific, but errors in amino acid selection are occasionally made, opening the door to inaccurate translation of the genetic code. The fidelity of protein synthesis is maintained by the editing activities of synthetases, which remove noncognate amino acids from tRNAs before they are delivered to the ribosome. Although editing has been described in numerous synthetases, the reaction mechanism is unknown. To define the mechanism of editing, phenylalanyl-tRNA synthetase was used to investigate different models for hydrolysis of the noncognate product Tyr-tRNA(Phe). Deprotonation of a water molecule by the highly conserved residue betaHis-265, as proposed for threonyl-tRNA synthetase, was excluded because replacement of this and neighboring residues had little effect on editing activity. Model building suggested that, instead of directly catalyzing hydrolysis, the role of the editing site is to discriminate and properly position noncognate substrate for nucleophilic attack by water. In agreement with this model, replacement of certain editing site residues abolished substrate specificity but only reduced the catalytic efficiency of hydrolysis 2- to 10-fold. In contrast, substitution of the 3'-OH group of tRNA(Phe) severely impaired editing and revealed an essential function for this group in hydrolysis. The phenylalanyl-tRNA synthetase editing mechanism is also applicable to threonyl-tRNA synthetase and provides a paradigm for synthetase editing.     

Differential distribution of D and L amino acids between the 2'' and 3'' positions of the AMP residue at the 3'' terminus of transfer ribonucleic acid.

Amino acids esterified to the ribose group of 5'-adenylic acid (AMP) constantly migrate between the 2' and 3' positions of the ribose at a rate of several times per second, which is slower than the rate of peptide-bond synthesis (15-20 per sec). Because the contemporary protein-synthesizing system only incorporates amino acids into protein when they are at the 3' position of the AMP at the terminus of tRNA, the value of the equilibrium constant relative to the 2' and 3' positions is of considerable interest. Differences between D and L isomers in this regard might be especially revealing. We have used N-acetylaminoacyl esters of AMP as models for the 3' terminus of tRNA and find that glycine and the L amino acids consistently distribute predominantly to the 3' position (approximately equal to 67% 3', approximately equal to 33% 2'), but D amino acids distribute to that position generally to a lesser extent and in a manner inversely related to the hydrophobicity of the amino acid side chain. This consistency of the L amino acid preference for the 3' position, combined with the inconsistency of the D amino acid preference, may be one reason for the origin of our contemporary protein-synthesizing system, which forms the peptide bond preferentially with L amino acids and only when they are in the 3' position of the ribose moiety of the AMP residue at the 3' terminus of every tRNA.     

Emergence of the universal genetic code imprinted in an RNA record

The molecular basis of the genetic code manifests itself in the interaction of the aminoacyl-tRNA synthetases and their cognate tRNAs. The fundamental biological question regarding these enzymes' role in the evolution of the genetic code remains open. Here we probe this question in a system in which the same tRNA species is aminoacylated by two unrelated synthetases. Should this tRNA possess major identity elements common to both enzymes, this would favor a scenario where the aminoacyl-tRNA synthetases evolved in the context of preestablished tRNA identity, i.e., after the universal genetic code emerged. An experimental system is provided by the recently discovered O-phosphoseryl-tRNA synthetase (SepRS), which acylates tRNA(Cys) with phosphoserine (Sep), and the well known cysteinyl-tRNA synthetase, which charges the same tRNA with cysteine. We determined the identity elements of Methanocaldococcus jannaschii tRNA(Cys) in the aminoacylation reaction for the two Methanococcus maripaludis synthetases SepRS (forming Sep-tRNA(Cys)) and cysteinyl-tRNA synthetase (forming Cys-tRNA(Cys)). The major elements, the discriminator base and the three anticodon bases, are shared by both tRNA synthetases. An evolutionary analysis of archaeal, bacterial, and eukaryotic tRNA(Cys) sequences predicted additional SepRS-specific minor identity elements (G37, A47, and A59) and suggested the dominance of vertical inheritance for tRNA(Cys) from a single common ancestor. Transplantation of the identified identity elements into the Escherichia coli tRNA(Gly) scaffold endowed facile phosphoserylation activity on the resulting chimera. Thus, tRNA(Cys) identity is an ancient RNA record that depicts the emergence of the universal genetic code before the evolution of the modern aminoacylation systems.     

From the Cover: CP1-dependent partitioning of pretransfer and posttransfer editing in leucyl-tRNA synthetase

Mistranslation is toxic to bacterial and mammalian cells and can lead to neurodegeneration in the mouse. Mistranslation is caused by the attachment of the wrong amino acid to a specific tRNA. Many aminoacyl-tRNA synthetases have an editing activity that deacylates the mischarged amino acid before capture by the elongation factor and transport to the ribosome. For class I tRNA synthetases, the editing activity is encoded by the CP1 domain, which is distinct from the active site for aminoacylation. What is not clear is whether the enzymes also have an editing activity that is separable from CP1. A point mutation in CP1 of class I leucyl-tRNA synthetase inactivates deacylase activity and produces misacylated tRNA. In contrast, although deletion of the entire CP1 domain also disabled the deacylase activity, the deletion-bearing enzyme produced no mischarged tRNA. Further investigation showed that a second tRNA-dependent activity prevented misacylation and is intrinsic to the active site for aminoacylation.     

Protein Synthesis Directed by Encephalomyocarditis Virus RNA: Properties of a Transfer RNA-Dependent System

Small amounts of encephalomyocarditis virus RNA direct a 50-fold increase in amino acid incorporation, in appropriately supplemented ascites tumor cell extracts, under conditions that give rise to authentic viral polypeptides. Incorporation in these crude extracts has a novel characteristic, namely, that it is almost entirely dependent upon the addition of exogenous tRNA. Further, this incorporation is restricted in that tRNA derived from ascites tumor cells or from rat liver permits translation of viral RNA, whereas tRNA from yeast or Escherichia coli does not. These translational barriers are due, at least in part, to an incompatibility between the tRNA of yeast and E. coli and the aminoacyl-tRNA synthetases of the ascites tumor cell. A more extensive basis for this incompatibility is suggested, however, by the failure of the E. coli aminoacyl-tRNA synthetases to restore viral RNA-directed protein synthesis in the presence of tRNA from E. coli, although the coli synthetases fully restore the poly(U)-directed synthesis of polyphenyl-alanine. The possible role that unique or favored codon classes might play in this restriction is considered, together with the implications of the observed requirement for tRNA.     

Protein synthesis editing by a DNA aptamer.

Potential errors in decoding genetic information are corrected by tRNA-dependent amino acid recognition processes manifested through editing reactions. One example is the rejection of difficult-to-discriminate misactivated amino acids by tRNA synthetases through hydrolytic reactions. Although several crystal structures of tRNA synthetases and synthetase-tRNA complexes exist, none of them have provided insight into the editing reactions. Other work suggested that editing required active amino acid acceptor hydroxyl groups at the 3' end of a tRNA effector. We describe here the isolation of a DNA aptamer that specifically induced hydrolysis of a misactivated amino acid bound to a tRNA synthetase. The aptamer had no effect on the stability of the correctly activated amino acid and was almost as efficient as the tRNA for inducing editing activity. The aptamer has no sequence similarity to that of the tRNA effector and cannot be folded into a tRNA-like structure. These and additional data show that active acceptor hydroxyl groups in a tRNA effector and a tRNA-like structure are not essential for editing. Thus, specific bases in a nucleic acid effector trigger the editing response.     

Evolution and relatedness in two aminoacyl-tRNA synthetase families.

Sequence segments of about 140 amino acids in length, each containing a selected consensus region, were used in alignments of the aminoacyl-tRNA synthetases with the aim of discerning their evolutionary relationships. In all cases tested, enzymes specific for the same amino acid from a variety of organisms grouped together, reinforcing the supposition that the aminoacyl-tRNA synthetases are very ancient enzymes that evolved to include the full complement of 20 amino acids long before the divergence leading to prokaryotes and eukaryotes. The enzymes are divided into two mutually exclusive groups that appear to have evolved from independent roots. Group I, for which two sequence segments were analyzed, contains the enzymes specific for glutamic acid, glutamine, tryptophan, tyrosine, valine, leucine, isoleucine, methionine, and arginine. Group II enzymes include those activating threonine, proline, serine, lysine, aspartic acid, asparagine, histidine, alanine, glycine, and phenylalanine. Both groups contain a spectrum of amino acid types, suggesting the possibility that each could have once supported an independent system for protein synthesis. Within each group, enzymes specific for chemically similar amino acids tend to cluster together, indicating that a major theme of synthetase evolution involved the adaptation of binding sites to accommodate related amino acids with subsequent specialization to a single amino acid. In a few cases, however, synthetases activating dissimilar amino acids are grouped together.     

Coding coenzyme handles: a hypothesis for the origin of the genetic code.

The coding coenzyme handle hypothesis suggests that useful coding preceded translation. Early adapters, the ancestors of present-day anticodons, were charged with amino acids acting as coenzymes of ribozymes in a metabolically complex RNA world. The ancestral aminoacyl-adapter synthetases could have been similar to present-day self-splicing tRNA introns. A codon-anticodon-discriminator base complex embedded in these synthetases could have played an important role in amino acid recognition. Extension of the genetic code proceeded through the take-over of nonsense codons by novel amino acids, related to already coded ones either through precursor-product relationship or physicochemical similarity. The hypothesis is open for experimental tests.