Logo IMG


Ode to the Code

Brian Hayes

Egged on by Error

Early guesses about the nature of the code often started from an assumption that it would maximize information density. One conjecture had each nucleotide base spelling out three messages at once. The concern with efficiency turned out to be misplaced; information density is not a very high priority for most organisms. The concept that has replaced efficiency as the great desideratum in genetic coding is error-­tolerance, or robustness. In one way or another, the code is thought to minimize the incidence and the consequences of errors in the transmission of genetic information, so that meaning can be recovered even from garbled messages.

Among the many ways that genetic signals could go awry, two kinds of errors have been singled out for attention: mistranslations and mutations. Errors in translation disrupt the reading of the genetic message—the flow of information from DNA to RNA and then to protein—but they leave the DNA itself intact. Translation errors were probably of great importance early in the history of life, when the machinery of protein synthesis was imprecise. Mistranslations are less frequent now, and less harmful. Each error disables only a single protein molecule. Mutations are another matter: They alter the DNA, the permanent genetic archive. Whereas a translation error is like an inkblot marring one copy of a book, a mutation is a flaw in the printing plate, reproduced in every copy. The simplest "point" mutations substitute one nucleotide for another at a single site on the DNA (with a corresponding change on the opposite strand).

The idea that fault tolerance might shape the genetic code arose as soon as biologists got their first glimpse of the codon table. The mapping from codons to amino acids is highly degenerate: In many cases multiple codons specify the same amino acid. But the synonymous codons are not just scattered haphazardly across the table; they clump together. Because of these clusters, a misreading or mutation has a better-than-average chance of producing a new codon that still translates into the same amino acid.

Closer examination of the table—with some knowledge of amino acid chemistry—revealed another possible strategy for coping with errors. When a change to a single nucleotide does not yield the same amino acid, it nonetheless has a good chance of producing one with similar properties. For example, all the codons with a middle nucleotide of U correspond to amino acids that are hydrophobic, or water-repellent, a trait governing how the chain of amino acids in a protein molecule folds up in the aqueous environment of the cell. Thus at least two-thirds of the time a point mutation in one of these codons will either leave the identity of the amino acid unchanged or will substitute another hydrophobic amino acid.

comments powered by Disqus


Subscribe to American Scientist