[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: D-amino acids in protein crystal structures

*** unsubscribe by sending the message "unsubscribe ccp4bb" to   ***
***       ccp4bb-requests@dl.ac.uk  or  Majordomo@dl.ac.uk       ***

It is not common at all for ordinary protein structures to contain D-amino
acids.  There ARE structures such as antibiotics and bacterial cell wall
components which legitimately contain D-amino acids, so such features are
not always errors.

It should be noted that TNT (which was used to refine these structures)
checks and reports chirality, but does NOT enforce it.  This was a
deliberate design decision.  There are a number of situations in which it is
difficult for a program to automatically move incorrectly placed atoms onto
the correct density.  In these cases the action of any automatic procedure
should be checked by the responsible crystallographer.  TNT permits but
flags incorrect chiral centers specifically to draw attention to areas which
require human evaluation.  This is in line with the general policy that if
you tell your refinement program everything you know about the structure you
have no way to test what comes out of the program.

The TNT Geometry module contains a SCREEN command which will give a list of
the worst outliers in EACH geometric category, and a similar command which
gives a list of the atoms with the largest structure factor gradients. 
These lists are a great help when heading to the graphics system for a final
structure review.  Other programs have similar tools.

Given that crystallography is an experimental science, and that the PDB
format contains no way to assign individual reliability estimates to the
atoms, the question of how to present results in sections of the structure
which are ambiguous is an interesting one.  If you let the blemishes show
through, the whole structure gets sneered at; if you make everything
plausible with no underlying data to support your position, you are lying to
the users of the coordinates.

Lynn Ten Eyck

colbc@wladziu.bio.purdue.edu wrote:
> *** unsubscribe by sending the message "unsubscribe ccp4bb" to   ***
> ***       ccp4bb-requests@dl.ac.uk  or  Majordomo@dl.ac.uk       ***
> How common is it for there to be D-amino acids in protein crystal
> structures? I have recently found two structures in the PDB (1DCO and
> 1DCP) by Endrizzi, Cronk, and Alber that have several D-amino acids.

Lynn F. Ten Eyck                      mailto:teneyckl@sdsc.edu
Science Department                    Tel: 619-534-5141
San Diego Supercomputer Center        Fax: 619-534-5152
P.O. Box 85608                        Holly Boyce: 619-534-8183
San Diego, CA 92186-5608              mailto:boyce@sdsc.edu