Naming of phosphate oxygens

Peter Slickers slickers at
Thu Nov 23 04:28:30 EST 1995

Henry Gabb wrote in response two my posting:

> Just out of curiosity, why do you want to distinguish between ... (the 
> two phosphate oxygens OP1 and OP2) ?

Because I want to calculate true rms distances between a crystral 
structure form the Brookhaven database and a modell. I have tried 
this for the drew-dickerson dodecamer '1BNA' using InsightII. 
In the structure build with Insight the two phosphate oxygens are 
pointing just to the opposite direction as in the crystal structure. 
If I now calculate the rms distance between the two structures I 
get a wrong value because the atoms in the two structures are 
aligned according to their ordering in the pdb files and _not_
according to their chiral positions.   

>> The structural meaning of this would be, whether OP1 is pointing towards
>> the minor or the major groove in a DNA double helix. 

> That suggestion works fine for a static structure.  Since the backbone
> torsions are flexible, the positions of the phosphate oxygens and C5'
> hydrogens with respect to the grooves can reverse.   
> ... 
> It's easy to draw a distinction between the H2' and H2'' atoms of the
> sugar; one will always be above the ring (i.e., on the same side as the
> C5') and one will always be below.

No, OP1 and OP2 will never reverse since P is a prochiral center and since the
position of P is fixed through O3' and O5'. With OP1 and OP2 it is
exactly the same kid of problem as with H2'1 and H2'2, but for H2'1/H2'2 
and H5'1/H5'2 the correct positions are defined by the IUPAC
recommendation of 1982 (IUPAC. 1983. Europ J Biochem, 131: 9-15.).
Therefore my question is, if there exist any recommendation for the 
phoshate oxygens.  


Peter Slickers                                  slickers at
Institut fuer Molekulare Biotechnologie
Beutenbergstrasse 11, Jena                      Tel.: +49-3641-65-6202
Postfach 10 08 13, D-07708 Jena, Germany        Fax:  +49-3641-65-6210

More information about the Molmodel mailing list