Early Evolution of Translation  - The Reverse Recognition Conjecture

1. Hypothesis on the origin of the genetic code:

Our contemporary triplet genetic code was preceded by an ancient doublet code (Wilhelm T and Nikolajewa S, 2004, Patel 2005,  Wu et al, 2005 , Copley et al, 2005 ). 
However, in order to avoid the frameshift problem one has to assume a triplet reading frame also in doublet coding times.
One still wonders about such  information wasting, where the third base would not carry any information at all.

2. Hypothesis on the early evolution of translation: early pre-tRNA (or tRNA precursor) binding was independent of the 5'-3' direction:

Figure 1.  Only two bases bound, either at the first two codon nucleotides, or at the last two ones. In this scenario all pre-mRNA nucleotides carry information, needed to translate the codon.

Figure 2. Two different tRNAs could recognize the same codon

The evidences to support the hypothesis of early translations
Order in the Genetic Code
Amino acids that have similar biochemical properties tend to have similar codons (The new classification scheme of the genetic code  and codon symmetry of the genetic code)
Special role of the second base of codons
  1. The amino acids with codon *U* ( that have U at second position)  are hydrophobic, whereas amino acids with codon *A* are hydrophilic (Taylor FJ and Coates , 1989).
  2. The codons of the form *U* define a set f five amino acids, all of which have very similar polar requirement, likewise *C* codons define nearly the same unique polar requirement. The codons CAY-CAR, AAY-AAR, and GAY-GAR each define a pair of amino acids (His-Glu, Asn-Lys, and Asp-Gly, respectively) that has a unique polar requirement. (Woese et al, 2000)
  3. The central triplet base the has the strongest interaction with the bases of 16S RNA (the universally conserved and essential bases A1492, A1493, and G530 (Ogle et al, 2001). Moreover, the middle base of the anticodon has particularly strong interactions with the correct aminoacyl-tRNA synthetase during amino acid attachment (McClain et al, 1998).
Evolutionary conserved group of amino acids
From analysis of the amino-acid replacement matrices, observation that reverse codon pairs (XYZ and ZYX codons) generally encode evolutionary similar amino acids (Thompson et al,1994) . We suppose that this observation is a relict from old ``reverse recognition times'', where the reverse recognition should have a minimal effect on the resulting polypeptide.

Strongly conserved groups of amino acids (Thompson et al, 1994) are subsets of exactly one quadrant, e.g. the amino acids M, I, L, V  belong to the upper right block in the new classification table of the genetic code. The other conserved strong groups belonging to one block of the new scheme  are MILF, STA, NEQK, NHQK, NDEQ, HY. The only exceptions are QHRK (R (Arg) is in another quadrant) and FYW (in three quadrants). In other words, reverse codon pairs tend to code for evolutionary similar amino acids, and each quadrant is enriched for amino acids with similar biochemical properties.
Observation about STOP codons
Reverse STOP codons (GAT, AGT, AAT) do not have their own tRNAs (just one exception in human).
Observation about tRNA
Ancient pre-tRNAs presumably only consisted of the anticodon loop, lacking the D- and T-loops (Rodin et al, 1993). Such pre-tRNAs would have been (almost) symmetrical and could thus bind in two directions. If the reverse recognition model is correct, the resulting polypeptide should be relatively independent of the pre-tRNA binding direction.
Observation about tRNA anticodons prevalence
 

1) Of course, no species has a tRNA with an anticodon complementary to any termination codon. Intriguingly, there is also no tRNA with an anticodon for a reverse STOP codon (GAT, AGT, AAT). The only exception is H. sapiens with one tRNAAsn with the anticodon ATT. The lack of specific tRNAs does not imply that no tRNA exists which can recognize reverse STOP codons. For instance, using base pairing allowed by Crick's wobble rules, tRNA with the GTT anticodon can recognize the reverse STOP codon AAT.

2). The significant suppression of  tRNAs with A at the first anticodon position. A** anticodons are fully excluded in archaea. In bacteria and eucaryotes there are some exceptions, but it can be observed that AY* anticodons do not appear in any species.

3). The significant suppression of A*A self-reverse codons. In no archaea and in no bacteria any tRNA has such an anticodon. In archaea the anticodon TAT is the only one without own tRNAs that is not a STOP anticodon or an A** anticodon. Interestingly, this is the only anticodon which according to Crick's wobble rules allows recognition of a STOP codon (TAG)