Re- Evolutionary tree of *all* proteins

biocukm at okway.okstate.edu biocukm at okway.okstate.edu
Sun Oct 16 10:38:44 EST 1994


In article <MAILQUEUE-101.941016104337.480 at molbiol.uct.ac.za>
ED at molbiol.uct.ac.za ("RYBICKI, ED") writes:
>> From:          biocukm at okway.okstate.edu (Ulrich Melcher)
>
>> I think we (Lartey, Voss and myself, ms in preparation) may have 
>come across a 
>> case where convergent sequence evolution has occurred during the 
>evolution of 
>> the ribgrass mosaic virus subgroup of tobamoviruses.  In their RNA 
>genomes the 
>> movement and coat protein cistrons overlap for 25 codons.  In 
>other 
>> tobamoviruses there is no overlap or a shorter overlap.  The 
>ancestral cistron 
>> in the overlap region can be deduced to be the coat protein one.  
>                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>
>Er...why?  Is it not possibly an N-terminal extension of the CP?  
>How "homologous" are those regions of the tobamovirus CPs?  Enough 
>to say that this has not occurred?

The evidence that the coat protein reading frame is ancestral is
as follows:
First, there is very little difficulty in aligning tobamoviral coat
 protein amino acid sequences.  Very few indels.  Second, analysis 
of the codon-position-dependent base compositions shows that the 
profile for positions 1, 2, and 3 in the coat protein overlap region
show the same preferences as do non-overlapping regions of the coat
and movement cistrons.  The exception, the T content, actually supports
the contention, since to derive the overlap, T's have to be suppressed
to avoid stop codons in the overlapping frame (Keese & Gibbs, 1992 
PNAS 89:9489).  Third, codon usage of the coat frame of the overlap
is closer to the usage in non-overlapping parts of both genes than is
the usage in the movement frame of the overlap.  
What measure of "homology" would you like?
>
>....I think that is what is meant by convergent 
>> sequence evolution.  The situation will certainly complicate 
>phylogenies based 
>> on movement protein amino acid sequences since all alignment 
>algorithms will put 
>> the C-termini together in some fashion, even though they are 
>encoded by non-
>> homologous nucleotides.
>
>This is definitely true IF it is the case...so, for all rules there 
>has to be an exception!
>
>  _________________________________________________________________
> [Rybicki sig deleted]

Ulrich Melcher
Dept Biochem Mol Biol
Oklahoma St. Univ.
Stillwater OK 74078
biocukm at okway.okstate.edu




More information about the Mol-evol mailing list