If the actual genetic code is to be found anywhere in living cells, it would be in the tRNAs and the enzymes that attach amino acids to them (tRNA synthetases). tRNA is not unique to humans:  tRNAs are essential for protein synthesis in all living things and may compose a significant portion of the number of an organism’s genes (they compose about 4% the number of human genes).

     tRNAs and tRNA synthetases are universal in living organisms and are essential for every human cell which makes protein.    Without them, human connective tissue cells could not make extracellular proteins,


the cells lining the kidney tubules of fish could not make the protein pumps for reabsorption and secretion,

and animal cells (such as the simple hydra below) could not make cell membrane proteins which allow the cells of a multicellular organism to interact with each other.
If the genetic code exists in any physical part of the cell, it is in the tRNA molecules and the enzymes which attach amino acids to them.  Some have proposed that the molecules of translation evolved in this order: tRNA, tRNA synthetases, precursors of the modern ribosome, the genetic code, mRNA, and finally proteins.  Although modern tRNAs bind their amino acids using protein enzymes, RNA molecules have been generated which can charge tRNAs with amino acids.  tmRNA is a RNA about 350 nucleotides long in bacteria which can serve as both mRNA and tRNA (Brosius, 2001). 

     The average tRNA is 80 nucleotides long.   The original transcripts of tRNAs may be twice as long but bases are removed and many bases are changed to form unusual bases (pseudouridine, inosine, methylguanosine, methylinosine).  Some of these modified bases are common to all tRNAs while others distinguish certain classes of tRNAs (e.g. wyosine bases are characteristic of tRNAphe).  Many nucleotide positions are relatively invariant among tRNAs.

      tRNA bases arrange to form an open cloverleaf structure (which is folded differently in a 3-dimensional view) with 3 loops (one orients the tRNA on a ribosome, one is an anticodon loop) and a region which binds amino acids.  The unusual bases are important in forming these loops.  There is an additional loop which is small in 75% of tRNAs (Class 1) but large in others (Class 2); its significance is not known. In a typical RNA, about half of the bases form double stranded helical regions with other parts of the molecule. These regions are clearly related to structure in tRNA and rRNA and mRNA must be unwound before it is translated (Gesteland, 1993).

      tRNA sequences are probably ancient, derived from the RNA world.  tRNA forms two hairpins.  The genes for tRNA possess an intron between these two domains, suggesting that the first cells might have had separate “minigenes” which coded for the separate halves of tRNA molecules.  At some point, these minigenes fused to from tRNA molecules.   If the assembly of minigenes into tRNA occurred after LUCA, then tRNAs coding for different amino acids in one species might be more similar to each other than to those coding for the corresponding amino acids in different species.  Sequence comparisons provides evidence which supports this (DiGiulio, 2001; (DiGiulio, 2004). The new Archaeal phylum Nanoarchaeota may represent the most primitive branch of life on earth and thus the “root” of the tree of life. The hyperthermophilic archaeon Nanoarchaeum equitans (which is a parasite of other archaea) uses split genes to code for at least 6 tRNAs and 11 proteins. tRNA molecules may have separate genes for their 5′ and 3′ halves in separate parts of the genome. Its rRNA sequences are unique. Unlike all other archaea, there are no operons in the genome (Di Giulio, 2007). As a result, tRNA may be a composite of separate genes combined after LUCA. Nanoarchaeum equitans may represent ‘Paleokaryotes’, surviving members of early cellular lineages which retain ancestral features (Di Giulio, 2006).

The top half of tRNA molecules are similar to molecules of modern genomes which are involved in the replication of genomes rather than translation.  It has been proposed that the bottom portion of the molecule (which includes the anticodon) was a later addition (Maizels, 1994).   tRNAs might have had primitive functions which preceded their incorporation into translation mechanisms.   For example, tRNAs might have been used to tag those ribozymes in precells which synthesized amino acids.  Some of the translation mechanisms (such as tRNA synthetases) might have evolved to improve the replication of ribozymes (Stevenson, 2002).

tRNA can perform other functions in certain organisms such as the transport of amino acids for cell wall construction, a feedback mechanism in the regulation of amino acid synthesis, and a transcription factor for RNA polymerase III (Gesteland, 1993).

      There are between one and two thousand tRNA genes in the human genome (perhaps around 1,300).  There are an estimated 50-60 individual genes (with multiple tRNA genes for certain amino acids obviously) and an average of 10-20 copies of each gene.  For example, there are 2 separate genes for methionine intiator tRNAs and separate genes for non-initiator methionine tRNAs; the human genone has an estimated 12 copies of the initiator methionine tRNA genes.  Multiple tRNA genes can be located together in clusters while other genes are solitary.  Some of the genes are pseudogenes.  The phosphoserine tRNA on chromosome 19q13 is interesting because it can override a stop codon.  It is the only natural opal suppressor known in higher eukaryotes—its TCA anticodon matches the UGA codon (which usually codes for stop).

     There are 22 genes for tRNAs in the human mitochondrial genome (including 2 copies of serine and leucine tRNAs).  Mutations in the majority of these genes have been linked to human disorders which include MERRF syndrome, MELAS syndrome, cardiomyopathy, SIDS, maternally transmitted diabetes-deafness, epilepsy, cerebellar ataxia, Parkinsons disease, and neonatal death.  One of these genes is the methionine tRNA homologous to the fmet initiator tRNA used in translation in prokaryotes but not in eukaryotes.

     SINE elements, which compose 13% of the human genome and 8% of the mouse genome are sequences which seem to have been derived from tRNAs (sharing an RNA polymerase III promoter) and amplified in the genome because of retroposition (Frenkel, 2004).

Given that there is a tRNA in archaea which can be charged with two different amino acids by two different tRNA synthetases, it is thought that tRNA synthetases evolved to complement a genetic code which already existed in the form of tRNA (Hohn, 2006).



    Although tRNA synthetases can occasionally attach the wrong amino acid to a tRNA, proofreading mechanisms reduce the error rate to about 1/60,000.  Typically, the enzyme does not change shape to add the amino acid when the incorrect tRNA is present.  After a tRNA has bound an amino acid, it is said to be charged and is referred to as amino-acyl tRNA.   tRNA requires the function of tRNA synthetases, which are proteins.  Since tRNA is required to make protein, there is the obvious question of how such a system could evolve.  Although in modern cells, the enzymes which attach amino acids to tRNAs are proteins, ribozymes have been isolated in vitro which can perform this aminoacylation.  Thus this important step for protein synthesis could have occurred without protein (Martinez Gimenez, 2002).  The tRNA synthetases are ancient proteins whose function was required in the establishment of the genetic code and may therefore have been one of the first proteins to evolve.  Interestingly, some of these synthetases may have performed various functions in early cells.  There are conserved regions shared between some class II aminoacyl-tRNA synthetases and the pol gB gene, a subunit of DNA polymerase (present in humans) (Carrodeguas, 2000).  

    Synthetases are actually a diverse group of enzymes which are not related to each other.  Each tRNA synthetase must distinguish one class of amino acid from the others.  They usually recognize at least one base in the anticodon and often the site that binds the amino acid is recognized.  One site near the amino acid attachment point is invariant in all isoaccepting tRNAs.  For example, arginyl-tRNA synthetase attaches the amino acid arginine to certain tRNAs; leucyl-tRNA synthetase attaches leucine to certain tRNAs; etc.  There apparently are a few exceptions to this rule: ProCysRS can attach the appropriate amino acids to different tRNAs; this dual-specificity enzyme is known from archea and simple eukaryotes (Bunjun, 2000).

     In prokaryotes, each tRNA synthetase functions separately.  In higher eukayotes, 11 synthetase polypeptides form a multienzyme complex.  Polypeptides may be results of ancient gene fusion: in both humans and Drosophila, glutamyl tRNA synthetase and prolyl tRNA synthetase (enzymes which are in 2 different classes) are joined in a single polypeptide chain (glutamyl-prolyl tRNA WARS1) (OMIM).

     The tRNA synthetases can be divided into two groups of 10 based on whether their ATP binding domain possess a Rossman fold (Class I) or a b sheet arrangement (Class II).  These two classes are unrelated to each other.  It seems that the ancestral members of these two groups were coded by the sense and antisense strands of the same ancestral gene.  Class II synthetases are related to those amino acids thought to have been used first in primitive proteins (ala, asn, asp, gly, his, lys, phe, pro, ser, and thr) and the class I synthetases with the amino acids which are thought to have been utilized later (arg, cys, gln, ile, leu, glu, lys, met, trp, and val) (Klipcan, 2004).  The two classes of tRNA may have originated as sense and antisense copies of the same ancestral gene (Carter, 2002). 

     Some have suggested that the charging of tRNAs with the appropriate amino acids arose at least twice and the current mechanisms are a fusion of two ancestral processes.  There are a couple of general rules observed in the charging of tRNAs.  First is the “class rule” in which all of the synthetases associated with any one amino acid are in the same class (either class I or II).  A second rule is the “monophyly rule” in which all of the synthetases associated with any one amino acid are more closely related to one another than any one is to the synthetase for any other amino acid.  A third rule is that all tRNAs which carry the same amino acid (and are “isoacceptors”) can be charged by a single synthetase enzyme.  These rules generally hold true; lysine is an exception to the “class rule” in that some organisms use a Class I synthetase to attach lysine while others use a Class II synthetase.  There are a few exceptions to the monophyly rule (Woese, 2000).  There are differences in the tRNA synthetases used across the domains of life. For example, glutamine may be attached to a tRNA by its own specific enzyme or by a relaxed enzymatic reaction for the attachment of glutamic acid to a different tRNA. After the Last Common Universal Ancestor, glutaminyl-tRNA synthetase and asparaginyl-tRNA synthetase were derived from glutamyl-tRNA synthetase and aspartyl-tRNA synthetase, respectively (Sheppard, 2008).

     Since synthetase enzymes perform equivalent roles in all organisms, there is a potential for horizontal transfer since the acquisition of a synthetase gene from another species would not be likely to be detrimental to an organism.  This seems to have happened given that some organisms have multiple equivalent synthetase genes in the same organism and since sequence comparisons using synthetase genes often produces very different results from those created through the comparisons of other genes (Woese, 2000).

     In the genetic code, the codon UGA usually signals “stop” but may,in both prokaryotes and eukaryotes (including humans),  be the site where selenocysteine is inserted.  There are a number of proteins which are needed for this to occur and two selenophosphate synthetase genes exist in humans (OMIM).  While most tRNAs and tRNA synthetases seem to have been present in LUCA, glutaminyl-tRNA synthetase is known only from eukaryotes and some eubacteria.  It seems to have arisen from a duplication of glutamyl-tRNA synthetase.   A similar duplication of a glutamyl tRNA sythetase which can charge a tRNA with glutamine rather than glutamic acid (Skouloubris, 2003).   Archeabacteria are more similar to eukaryotes than prokaryotes in their phenylalanine and methionine synthetase (Woese, 2000).

     Genomes studied to the present not only contain multiple copies of tRNA synthetases, they can include modified copies of these enzymes which have been recruited to other functions such as the synthesis of amino acids, DNA replication, and the regulation of translation (Ibba, 2004; De Farias, 2008). tRNA synthetases have been modified to perform functions completely unrelated to the charging of tRNAs. E.coli possesses a modified tRNA synthetase modifies RNA to created the nucleoside queuosine, other bacteria possess an enzyme which functions in the synthesis of histidine, and archaea possess a modified enzyme which synthesizes arginine (Sissler, 1999; Roy, 2003; Salazar, 2004).

Archaea and eukaryotes use the same enzymes, SerRS, PSTK, and SepSecS, to convert the amino acid phoshoserine to selenocysteine (Sec). It is the only one of the 22 amino acids to lack its own tRNA synthetase. Instead of being directly attached to tRNAs, selenocysteine is produced by modifying other amino acids (Yuan, 2006).