IUBio GIL .. BIOSCI/Bionet News .. Biosequences .. Software .. FTP

Mg2+ Sigma value

Barani barani at mace.cc.purdue.edu
Thu Mar 14 09:52:11 EST 1996

>Article: 374 of bionet.software.x-plor
>From: J.Martin at MAILBOX.UQ.OZ.AU (Jenny Martin)
>Subject: Mg2+ parameters
>Date: 12 Mar 1996 14:38:33 -0800
>Does anyone have any suggestions for parameters for Mg2+ using  the
>parhcsdx.pro forcefield?  We have been trying the parameters from
>parallhdg.dna  (eps 0.2 and sigma 2.8509) but this does not seem to work
>very well for a Mg bound at an active site and surrounded by waters. The Mg
>keeps getting refined out of density and the refined water-Mg distances are
>too long. Also, should a charge of +2 be used on the Mg or should it be
>kept neutral?  Is there anything else that needs to be defined?
>Any suggestions or comments would be very gratefully received!
>Thanks very much,
>Jenny Martin

 If I am not wrong, the sigma value 2.8509 is VDW radius only 
 for oxygen. However, the radius of magnesium is only around
 1.35.  The gives (Sigma1+Sigma2) = 4.2 and the distance between
 water and magnesium is half this value = 2.1. The distance 2.1
 corresponds to a charge of 0.33 (based on a theorem by Brown) 
 and if there are six ligands between magnesium and oxygens, the 
 total charge equals 2.0. 

 Well, the above rationality is what I used and I obtained reasonable
 distances close to 2.1. If the sigma values are correctly chosen
 there is no need to be concerned about the +2 charge. However,
 I too had the same difficulty of letting magnesium slip away from
 its expected position in the map. Hence, it is better to either fix
 or restraint the metal. In your case, the distances are too large,
 hence I would recommend not to turn off the electrostatic term by
 making the charge zero. In my case the distances were too short
 and I had to turn off the charge on metal only.

 I recommend a Sigma value for magnesium ranges between 1.2 and 1.4.


 chumma at purdue.edu

More information about the X-plor mailing list

Send comments to us at archive@iubioarchive.bio.net