In article <41kpkj$b3k at newsbf02.news.aol.com>, hpyockey at aol.com (HPYockey) writes:
>>Any knowledge of coding theory or practice shows that a change in the
>sense assignment of a letter in the sending alphabet without a
>corresponding change in the receiving alphabet is effectively noise. It is
>worse than random noise because it makes an error whenever that letter of
>the sending alphabet appears.
some stuff deleted
>This process is irreversible, since all backward steps increase the
>vulnerability to mutations to non-sense codons. For the same reason the
>paradigm assumes that once an assignment is made it does not change."
This argument appears to make sense; fortunately evolutionary change
can be more circuitous- there is a way around this.
If I may paraphrase the argument above:
Once a full genetic code has been developed and is used (all codons
used in the genome for instance) then a change in the assignment
of one of the codons will wreak havoc in the genome- for instance
coding UGA as tryptophan is going to screw up the expression of
numerous genes which use UGA as a stop codon. Thus any change in
in the assignments of codons is likely to be selectively deleterious
in the extreme and CAN never happen.
There is a flaw in this argument. Firstly it is not NECESSARY in
any sense that all 64 codons are used in a genome. So
if the codon UGA were not present in a genome at all, then a change
in the assignment of this codon will NOT HAVE ANY EFFECT AT
ALL on the organism's fitness. Thus a mechanism for codon
reassignment could be the following (suggested by Osawa and Jukes
many times):
1) Codon bias becomes very extreme and a certain codon (call
it codon A coding for amino acid Z) is lost in a genome (never used)
2) Gene duplications generate tRNA genes (coding for amino acid
X) with changes so that the anticodons are complementary to the
above lost codon A
3) Codons in the genome coding for X, are now free to mutate to
codon A, because there is a tRNA which will correctly read them.
In this way codon A has been reassigned from coding for X to
coding for Z.
There is no selectively deleterious intermediate in this scenario.
Note that evidence for such a codon reassignment would be
extreme codon bias against using codon A. Such a bias is
noticed in some of the examples of codon reassignment that have bee
found so far.
Cheers
Andrew Roger
aroger at ac.dal.ca >
