Volume 1

Gene expression / Benjamin Lewin.

  • Lewin, Benjamin.
Date:
1974-1977
Catalogue details

Licence: Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)

Credit: Gene expression / Benjamin Lewin. Source: Wellcome Collection.

    58/668 (page 36)
    in which the more error-prone codons represent amino acids which are functionally less important to proteins. Amino acids which tend to be involved in the mechanisms of enzyme catalysis—those with polar side chains—would tend to be assigned to the least error-prone codons. Also, codons which are likely to be mistaken for each other during translation would tend to be assigned to the same, or at least to functionally related, amino acids. Woese observed that the probabilities that misreading will occur at each of the three nucleotides in a codon are 100:10:1 from position three:one:two. Position three is the most degenerate in the genetic code, so that codons differing in this base position tend to be assigned to related amino acids. The second base is the least error-prone and probably the least involved in such restraints. As evolution progressed, the cell would make more precise distinctions amongst amino acids and begin to recognise individual species rather than just functionally similar groups. Selection would favour the development of a less error-prone and more complex translation apparatus; and only when the codon catalogue has become (virtually) completely ordered in its amino acids would it be possible to evolve a really efficient translation apparatus. (The ways in which this might happen have been discussed by Woese, 1967 and Orgel, 1968). Any cell succeeding in this development would have such an enormous selective advantage that all living cells would be descendents of this single cell line. As translation evolves into a more sophisticated form, the effect of changing any particular codon assignment increases greatly; eventually, such change becomes lethal and is no longer possible. Evolution of the code itself must therefore have been confined to the earlier and more fluid stage of codon assignments. Crick (1968) has suggested that the code might start in a primitive form in which a small number of triplets coded for comparatively few amino acids. The assignment of codons to amino acids would be random—so this is a purely selective model. Crick argued that the code is most likely to have started as a triplet one, since if it commenced with some other coding ratio, all previous messages would become nonsense, and presumably lethal to the cell, with the switch to triplet reading. (This argument disposes of all models which suppose that the code might have started as, say, sixteen dinucleotides coding for few amino acids, later expanding to a triplet system coding more amino acids.) It is not clear just which amino acids would be present in this initial code, although some of the more complex species present today are more unlikely to have existed under prebiotic conditions. In an intermediate phase, these primitive amino acids would take over most of the triplets of the code so as to reduce to a minimum the number of codons without meaning; the extension of codons for any particular amino acid would probably be to those related to the currently assigned codons. The final code would form as new amino acids replaced some of the primitive ones. This would happen most easily if the amino acid was related to that previously coded by