How much homology is needed to function?

Michael Holloway mhollowa at ic.sunysb.edu
Sat May 29 16:53:34 EST 1993


In article <16BDA853C.SHICKLEY at VM.TEMPLE.EDU> SHICKLEY at VM.TEMPLE.EDU writes:
>Organization: Temple University
>X-Newsreader: NNR/VM S_1.3.2

>My question is: Has anyone else seen evidence of small regions of
>homology which have "actual" functional significance? How much
>conservation is necessary to carry out a job for a protein? We
>have some evidence that regional homologies, albeit small, can
>confer suprising proerties on a protein--if one looks for them.

We just had a symposium on protein-protein interactions in transcriptional
interactions here and comparing homologies of helix-loop-helix domains 
came up several times during different talks.  In comparing the HLA domains 
of POU-homeo domains, aligning the alpha helixes was needed before the 
homology of three residues could be seen (if I recall correctly).  In other
words, you had to have an hypothesis about the function of putative tertiary
structure before you could see any homology.  The homology is just too small 
to have an algorithm just spit it out at you.  Winship Herr was comparing 
the lambda repressor with the POU-homeo domains of Oct1 and 2 and with 
Herpes VP-16.

Mike



More information about the Bioforum mailing list