Dear Crystallographers,

Having looked into the structure I mentioned using the "atomic contacts" feature of the whatif server, it seems that both the 1.0 Ang structure I previously mentioned, as well as another, lower-resolution related structure, have a significant number of "waters" which are < 2.6 Ang from polar atoms, with many in the 2.4 Ang regime (suggesting Na+). Both structures were solved in the presence of Na+. It seems impossible for water to bind at this distance, so would it not be best for the PDB to add some distance cutoff to its validation process? Should the PDB retrospectively review all structures for "waters" binding at these distances, cross-checking, of course, for sodium in the protein stock and/or crystallization condition? I am not sure about the profundity of the consequences of such a natrification, but truth is truth, however mundane, no?

Regards,

Jacob Keller

ps I am curious about how many structures would be affected by this.

*******************************************
Jacob Pearson Keller
Northwestern University
Medical Scientist Training Program
Dallos Laboratory
F. Searle 1-240
2240 Campus Drive
Evanston IL 60208
lab: 847.491.2438
cel: 773.608.9185
email: [log in to unmask]
*******************************************

----- Original Message ----- From: "George M. Sheldrick" <[log in to unmask]>

To: <[log in to unmask]>

Sent: Tuesday, February 16, 2010 5:00 AM
Subject: Re: [ccp4bb] Distinguishing Between Na+ and H2O

A few years ago we thought that the bond valence method might provide
an answer to this problem and wrote a paper on the subject (Acta Cryst.
2003 D59 32-37). Subsequent experience has convinced me that although
this method works well for identifying ions such as Mg2+ and Ca2+ with
good resolution data, it is not reliable in other cases such as Na+,
and for this reason I never distributed the version of SHELXPRO that
includes the bond valence test (I would not like my programs to get a
bad name).

In fact we currently have a protein crystallized from a high NaCl
concentration that is giving us a lot of problems distinguishing
between Na+ and water molecules. Since we were able to find some
chlorides in the anomalous map we know that cations must also be
present, but the anomalous data are too weak to help much with Na+
because of its lower f".

There are however some tentative indicators for Na+. The bond valence
sum (the sum of the 'bond orders' to the surrounding atoms estimated
from the distances) tends to be higher than for Cl- or H2O.
Tetrahedral coordination is more likely to be water or Cl-, Na+
prefers 5 or 6 neighbors. And of course two cations (or two anions)
that are close to each other should not have an occupancy sum greater
than unity.

George

Prof. George M. Sheldrick FRS
Dept. Structural Chemistry,
University of Goettingen,
Tammannstr. 4,
D37077 Goettingen, Germany
Tel. +49-551-39-3021 or -3068
Fax. +49-551-39-22582


On Tue, 16 Feb 2010, Eleanor Dodson wrote:

Yes - we are puzzling over the same phenomena.

Look at this web site set up by Marjorie Harding

http://tanna.bch.ed.ac.uk/

It lists the likely coordination patterns.

We certainly have ideal Na bonding in our structure - but unfortunately we
wanted to find Ca which has a very similar pattern!!

Grrr

Eleanor






Jacob Keller wrote:
> Dear Crystallographers,
>
> I am looking at a 1.0 Angstrom structure which contains many waters, > but I
> am wondering whether some of them might really be sodium ions. Is there > any
> straightforward way to distinguish between these two, and if so, what > is the
> software which implements this? Although the electron density > difference
> between sodium and water should be very small, perhaps the binding > geometry
> would provide a clearer distinction? Has anybody encountered this > question
> before?
>
> Regards,
>
> Jacob Keller
>
> *******************************************
> Jacob Pearson Keller
> Northwestern University
> Medical Scientist Training Program
> Dallos Laboratory
> F. Searle 1-240
> 2240 Campus Drive
> Evanston IL 60208
> lab: 847.491.2438
> cel: 773.608.9185
> email: [log in to unmask]
> *******************************************